The Blog
Articles on Wealth Building, Business & Real Estate
Insights from the author of The Trap Series — on escaping the rat race, starting a business with no money, avoiding bad real estate, and building lasting wealth.
Automated Hosting & Indexing Health Checks: Fix Google Search Console Errors Before They Tank Your Rankings
Detect and fix common Google Search Console errors including redirects, canonicals, duplicate content, and HSTS with the automated Hosting & Indexing Health tool.
Static Site Generators Compared: Eleventy vs Next.js vs Astro vs Hugo in 2026
Head-to-head comparison of Eleventy, Next.js, Astro, Hugo, Gatsby, and more with build benchmarks, learning curves, and a decision matrix for your next project.
NeonDB: Serverless Postgres for Side Projects and Scaling
Learn when to use NeonDB for serverless Postgres, how branching works, connecting from Netlify and Vercel functions, and free tier limits explained.
Review Velocity for Local Businesses: The Cadence, Response Targets, and Aggregators That Actually Move Rankings
Google, Yelp, and Apple Business all weight recency and response time. Here's the cadence to hit by industry, the response SLAs that hold rank, and which aggregator tier to buy by scale.
Three Direct-Booking Patterns That Don't Break OTA Rate Parity (Without Getting Delisted)
Booking.com and Expedia contracts punish hotels that advertise a lower public rate. Here are three incentive patterns that pull direct bookings without breaching the parity clause, plus one that looks fine but isn't.
Supabase with OAuth: When You Need It and How to Set It Up
Learn when your project needs Supabase OAuth authentication and follow step-by-step setup for Google, GitHub, and email login with Row-Level Security.
Competitive Social Media Analysis: Generate AI Prompts to Clone Winning Strategies
Use the Social Media Content Prompt Creator to reverse-engineer competitor strategies for Instagram, TikTok, Medium, and Meta with AI-generated content plans.
How to Build AI-Powered SERP Analysis Prompts for Any Website
Use the free SERP Analyzer Prompt Generator to create AI prompts that audit keyword density, content gaps, and title optimization from live search results.
Generate AI Fix Prompts From Any Site Audit: Let ChatGPT or Claude Write Your Code
Turn site audit results into ready-to-paste AI prompts that generate exact code fixes for SEO, schema, performance, and accessibility issues.
How to Compare 10 Websites at Once: The Batch Analyzer for SEO, GEO, and AEO
Use the free Batch Analyzer to compare up to 10 websites across 8 scoring dimensions — SEO, GEO, AEO, schema, performance, and more — with exportable reports and AI fix prompts.
How to Trigger a Google Knowledge Panel: The E-E-A-T Endgame (Part 5 of 5)
The exact combination of Wikidata, schema markup, external profiles, and convergent signals that triggers Google to create a Knowledge Panel for your personal or brand entity.
E-E-A-T Reviews & Social Proof: How to Build Trust Signals Google Can Verify (Part 4 of 5)
How to collect reviews, implement AggregateRating schema, leverage testimonials, and build the social proof layer that completes Google's trust evaluation of your brand.
One Public Pension Is 25% Funded and Another Is 103%. The Difference Is Not What You Think
Chicago's fire pension is 25.25% funded. Colorado's is 103.5%. After reading 36 valuations across six states, three things predict which end you land on.
E-E-A-T Schema Markup: The JSON-LD That Makes Google Understand Your Brand (Part 3 of 5)
How to implement Person, Organization, Book, and BookSeries schema.org markup with sameAs, knowsAbout, and entity cross-references. The exact JSON-LD blocks that feed Google's Knowledge Graph.
Idaho cut school property tax by $35 million in one year, and paid for it with sales tax
Idaho delivered $328.0 million of property tax relief in tax year 2025 and school levies fell $34,975,174. The cost did not vanish. It moved to the general fund.
Four of Six Colorado Municipalities Have No Fire Department, and the Fire District Charges More Than the Town Does
Parker levies 2.602 mills. South Metro Fire Rescue levies 12.250 on the same homes, 4.7 times as much, and none of it sits in any city budget. Breckenridge: plus 47%.
Idaho Falls Power moved $4,347,810 out of the utility fund in FY2025. It is on no one's tax bill.
Idaho Falls Power transferred $4,347,810 out of the utility fund in audited FY2025, up 16.8% in two years. It is a charge on ratepayers, outside the 3% cap.
Twin Falls Has No General Obligation Debt and a 0% Health Premium Increase. Police and Fire Are Still 51.70% of the General Fund.
Twin Falls carries no general obligation bonds, holds a Moody's Aa2, and budgeted a 0% health premium increase for FY2027. Police and fire still take 51.70%.
Boise Raises 62.1% of Its General Fund From Property Tax. Its Levy Rate Just Fell and the Bill Still Went Up
Boise's FY2027 proposed budget takes $216,007,833 of property tax into a $347,821,431 general fund. Parker, Colorado takes 4.4%. The cap explains the gap.
Breckenridge Budgets 3.9 Times More for Workforce Housing Than for Police
Breckenridge's 2026 adopted budget puts $22,863,450 into workforce housing and $5,801,816 into police, 17.3% of the general fund, the lowest share in a ten city study.
The API Refused Almost Every Request. The Same Data Was Sitting in a Free Bulk File.
A feed that covered 2 of 7 topics now covers 7 of 7, in about four seconds, with no new infrastructure. The fix was reading the publisher's own bulk files instead of its rationed query API. Here is how to tell which kind of failure you actually have.
Centennial has no pension liability at all. Its sheriff's contract still rose 27.5% in three years.
Centennial has no net pension liability and no OPEB in its audited statements. Its sheriff's contract still rose 27.5% and now takes 58.9% of general fund operations.
Building Your E-E-A-T Profile Network: Google, Wikidata, LinkedIn & More (Part 2 of 5)
Step-by-step guide to creating authority profiles on the seven platforms that feed Google's Knowledge Graph. Wikidata, Google Business Profile, LinkedIn, Crunchbase, Amazon Author Central, ORCID, and how they connect.
Greenwood Village has levied 2.932 mills every year since 2016. Its residents' tax bills went up anyway.
Greenwood Village's municipal mill levy has been 2.932 since 2016, about $185 per $1 million of value. The fire district that serves it just added 3.000 mills.
Parker, Colorado has a $0 pension liability and collects 3.1% of your property tax bill
Parker's share of the FPPA net pension liability is $0, almost 80% of its retirement spending is defined contribution, and the town is 3.1% of the property tax bill.
Aurora's police and fire already take 58.8% of general fund operating spending, and took 61.8% of the growth
Aurora's 2026 adopted budget puts police and fire at 58.84% of general fund operating spending, up from 58.30% in 2023, taking 61.8% of the growth.
Denver's own pension plan is 59.97% funded, and the general fund paying for it is shrinking
Denver's employee retirement plan is 59.97% funded with a $1.62 billion net pension liability, and the general fund carrying it fell 6.2% in two years.
ETO Scout: 5,953 Chinese tech articles summarised in English, free, and deliberately not a dataset
Scout reads 13 Chinese-language tech sources and summarises them in English, free. It also publishes a blunt list of things you should not use it for. That list matters.
The City With No Pension Liability Is the One Getting Squeezed Hardest
Centennial carries no net pension liability at all. Its police contract still rose 27.5% in three years and took 70.2% of general fund growth.
OpenAlex: 324 million papers given away free, run by a nonprofit that spends $4.2 million a year
OpenAlex indexes 323,964,135 works under CC0 and rivals Scopus and Web of Science. The API just moved to metered pricing. What it costs, and what it gets wrong.
Two States, Two Budgets, One Very Different Bill: Colorado and Idaho in 2026
Colorado PERA is 75.31% funded with a $24.19 billion shortfall. Idaho PERSI is 90.89% while assuming lower returns, so the gap is wider than it looks.
The Official Count Says 5,234 Laundromats. The Real Number Is 20,293. How to Size Up Any Local Market With Public Data
The most-quoted federal business counts miss most small operators, understating one industry by 3.9x. Here is the free-data method that finds genuinely underserved trade areas, and the traps that make it lie to you.
A Study Like This Quotes at $15,000. It Ran in an Afternoon. The Cheap Part Is Not the Interesting Part.
A two-metro site-selection study prices at $15,000 to $35,000 and takes about 90 hours. The same work ran in an afternoon. The real finding is that the expensive version usually ships a number that is wrong by 3.9x.
Harvard Dataverse: 302,000 datasets, no account needed, and a 2.5 GB ceiling nobody tells you about
Harvard Dataverse indexes 302,139 datasets and has served 161 million downloads. Here is what it costs, where the free tier caps, and how to get data out in three commands.
E-E-A-T Authority for Brands: What It Is and Why Google Demands It (Part 1 of 5)
Google's E-E-A-T framework — Experience, Expertise, Authoritativeness, Trustworthiness — determines which brands rank. Here's what it actually means and how to build it from scratch.
The Biggest Deduction on Your Pay Isn't a Tax — and It Never Appears on Your Pay Stub
A private employer spends $3.41 an hour on health insurance and $2.78 on Social Security and Medicare combined. Family premiums grew 2.4x in real terms since 1999 while wages grew 1.22x.
One Resume Per Job Is Now the Standard, and Not for the Reason You Think
LinkedIn now takes roughly 11,000 applications a minute. Workday told a federal court its software rejected 1.1 billion of them. Here is what actually screens you out, and a free prompt that builds a tailored resume without inventing anything.
I Read All 500 Pages of the 2026 Trustees Reports So You Don't Have To — Here Are Four Infographics
Social Security's retirement fund empties in 2032, not 2034 — the later date assumes a law that does not exist. Four sourced infographics, free PDFs, every figure with its table and page number.
GDELT: the quarter-billion-record open dataset that lets a small newsroom watch the whole world
GDELT indexes news from over 100 languages every 15 minutes and gives it away free. Here is who built it, who pays for it, and what it took to put it on a working news site.
The listening layer: ASR, Google Chirp, Jigsaw and the programs holding up global media monitoring
A quarter century of world television was machine translated for $54,000. Here is the speech infrastructure that made it possible, who funds it, and which pieces are quietly being switched off.
Everyone Fights Fire. Almost Nobody Prevents It. What Portugal, Australia and the EU Do Differently
In 2023 the US spent $3.2 billion fighting fire on the fewest acres since 1998. Portugal moved prevention from 20 percent of its wildfire budget to 46 percent in four years. Read both ledgers.
Veterans, SDVOSB Certification, and the Free Advisor Network Almost Nobody Uses
Certification moved from the VA to the SBA in January 2023, and a free nationwide advisor network got renamed in the same window. Here is what a veteran founder should actually do, in order.
Together AI for a Small Business: What It Actually Costs, What Works, and Where It Is the Wrong Tool
Verified Together AI pricing, working code, and the two findings that change the retrieval advice: there is exactly one serverless embedding model with a 514-token context, and rerank is not on serverless at all.
One Vote a Share: How the Tisch Family Keeps Loews by Making It Smaller
Loews has no supervoting stock and no holding company above it. The Tisch cousins hold fewer shares between them than in 2018 and a far larger slice of the company. Here is why.
Forty Percent Is the Number That Matters: How Two Charitable Trusts Hold Tata
Tata Sons' Articles do not key control to the trusts' 65 percent. They key it to 40. At that line two trusts still name a third of the board and the chairman.
What Hosting a New Business Site Actually Costs in 2026: Cloudflare, Netlify, and Vercel, With the Real Numbers
Netlify quietly added a $9 tier, Cloudflare bills static asset requests at zero, and Vercel charges per seat. The current numbers, what actually generates a bill, and which one a launching business should pick.
The Pre-Launch Audit Loop: Eight Things to Check Before You Tell Anyone the Site Exists
Most new sites launch with broken email authentication, no structured data, and an accessibility problem nobody looked for. Here is the eight-dimension check that finds all three in an afternoon.
Observe, Evaluate, Launch: The Whole Playbook for Starting a Brand Without Hiring Anyone
Name clearance, entity, EIN, site, schema, listings, and the free federal advisors, in the order they actually have to happen. Every step cited, every tool free, and a downloadable URL map.
Mintlify's Data Page Says Two Thirds of Documentation Traffic Is Now Agents. Here Is What That Costs You.
Mintlify publishes a live report showing agent traffic at roughly 65% of requests across every docs site it hosts, and 82.5% in crypto. Here is what the page actually is, what the platform costs, and the free GitHub alternatives that hold up.
Ten to One, Moved Downstairs: How Marriott Family Control Survived Losing the Dual-Class Vote
Marriott shareholders voted the family's supervoting stock out of existence in 1998, over the board's objection. The ten-to-one ratio moved into a private Delaware box.
Zero Tax, Full Collateral: How Offshore ILS Vehicles Underwrite American Catastrophe Risk
Cayman and Bermuda vehicles hold $65.8 billion of catastrophe risk without US insurance licenses. The collateral rule meant to keep them out is what lets them in.
Nobody Is Entitled to IKEA: Two Foundations, a 3 Percent Fee, and a Family Kept in the Minority
IKEA's two halves are owned by foundations legally barred from paying any beneficiary. The Kamprad heirs inherited board seats, not shares. Here is the structure, from the filings.
The Escape Hatch in the Will: How the Hilton Fortune Went to Charity and the Company Went to Blackstone
Conrad Hilton's will sent more than 99 percent of his estate to charity and, in the same paragraph, gave his son an option to buy it back. That clause took ten years to settle.
The Free US Small Business Help You Are Already Paying For, and Which Door to Knock On
Four federal advisor networks, a certification portal, and a beneficial-ownership rule that reversed in 2025. A plain map of the free help available to a US small business, and the paid impostors to avoid.
Every Place Your Business Needs a URL, and the One Line of Schema That Ties Them Together
Google, Bing, Apple, Yelp and Nextdoor listings are all free and take an afternoon. The part almost everyone skips is wiring them back to one entity so a machine knows they are the same business.
The AI Quick Start for Launching a Business: What to Install, What to Pay For, and What You Will Not Own
One subscription and a terminal will do the market research, write the site, generate the brand assets, and run the audits. Here is the stack, the current prices, and the copyright limit nobody mentions.
How to Diversify Your AI API Spend Across Providers Before Your Vendor Cuts You Off
Most advice about escaping AI vendor lock-in makes your bill worse. The Anthropic batch discount does not exist on Bedrock or Vertex, and your likeliest cutoff is your own tier spend cap. Here is what actually works, with verified top-up URLs.
List Your Books on Google Shopping for Free via Merchant Center
Google Merchant Center lets you list physical and digital products in Google Shopping results for free. Here's how to set it up for books without having a traditional e-commerce store, including direct links to Amazon.
The Chapter 35 Email Going Around Is Right About the Money and Wrong About Who Qualifies
A forwarded email says the age limit is gone if the veteran's P&T date is after August 1, 2023. The statute keys off the child, not the rating. Families are ruling themselves out wrongly.
Trademark Your Logo With the USPTO: What It Actually Costs and How to Lock the IP Down
The USPTO charges $350 per class to file. Only about a third of applications clear examination without an office action. Here is the real cost and the real process.
LinkedIn Newsletters: 100% Reach to Every Subscriber (Not 5% Like Posts)
LinkedIn newsletters deliver to every subscriber's inbox and LinkedIn notification feed — bypassing the algorithm that throttles regular posts to 5% reach. Here's how to set one up, what to write, and how it cross-promotes your books and sites.
One Account, Twenty Four AI Providers: How a Small Business Actually Uses Cloudflare AI Gateway
AI Gateway is free on every Cloudflare plan and creates itself on your first request. Here are the request shapes that work, the header that causes most 401s, and what Code Mode really saves.
Audit Your Own Web App's Security Before You Pay Anyone: The Two-Phase AI Method
A pentest quote runs $5,000; a SaaS scanner charges $99/mo and keeps your findings. Audit your own app, WordPress store, and server exposure yourself, free: the two-phase method and one master prompt.
Vinson Hall and the Armed Forces Retirement Home Are Not the Same
Vinson Hall is a private nonprofit community with roots among officers. The Armed Forces Retirement Home is a federal, enlisted oriented home. Same mission, no affiliation.
Silent Partner in a Small Deal? What a Real Offer Looks Like Before You Wire Money
A friend weighed a small passive stake in a gym and asked if it was fair. Here is how real private real estate deals are structured and the one test that kills the bad ones.
The real tax picture for tribal members and nonmembers in Indian Country
Living on a reservation is not a tax haven. A nonmember gets no personal income tax break, so the real leverage is business structure and aligning with the tribe.
DMARCbis Is Here: What the New DMARC Standard Actually Changes for Your Domain
DMARC was re-published in May 2026 as RFCs 9989, 9990, and 9991. Your existing records still work. Here is what actually changed, and what to do about it.
Who Owns the Firehawk: The Colorado Shop That Turns Black Hawks Into Firefighters, and the Bankruptcy Above It
United Rotorcraft builds the Firehawk that Cal Fire and Colorado fly. Its owner is an air-ambulance company whose 2023 bankruptcy wiped out its private-equity backers.
When the Network Owns the Lender: Credible, Yrefy, and the Financial Pitch Woven Into Cable News
The mortgage-rate news on Fox Business is written by a lender Fox owns. Yrefy paid TV hosts to endorse its notes and was fined $750,000 for not disclosing it.
CLI or Browser? How the Terminal, the Web Chat, and the Cloud Agent Really Differ on Speed, Cost, and Who Holds Your Code
The real speed gap is not terminal versus browser, it is whether the tool closes its own loop. Eight AI coding stacks compared on speed, cost, and where your code goes.
Two concepts, one kitchen: putting a day brand and a night brand online without confusing Google
One kitchen, two day-part brands, one address. A second Google profile can get both suspended, and a subdomain can bury the new concept. Here is the playbook that works.
Eight data-flow security habits from running real pipelines (Part 5 of 5)
SSRF on fetches, auth on every endpoint, running equals committed, one writer per dataset, and knowing what data leaves the box. Eight habits I picked up the hard way running data pipelines.
Why Add a PWA to Your Site: The Honest Value Case for a Progressive Web App
A progressive web app makes your existing site installable, offline-capable, and push-enabled with no app-store cut. Here is the honest value case, the real numbers, and the iOS caveats.
The Toll Booth: The Index of Who Collects Rent on the Things You Cannot Skip
A running index of the toll economy: the middlemen a law makes unavoidable, the fees baked into prices you never itemize, and the rent collected on money you already own.
How Members of Congress Are Paid, Pensioned, and Insured: The System an Outsider Joins
A bartender in 2018, a member of the House in 2019. The salary is frozen and losing to inflation, but the pension is indexed and the government pays about 72 percent of the health premium.
DigiGone: Who Owns the Encrypted Doctor at Sea, and Who Benefits
DigiGone's encrypted telemedicine kits run on standard AES, a wall of trademarks, and a veteran-owned federal set-aside. Here is who owns the doctor in the box at sea, and who benefits.
Run a Design System Generator on Your Own Site: The UI UX Pro Max Skill
A local, database-driven AI skill that recommends a page pattern, style, color palette, font pairing, and the UX rules to follow, generated for your exact product in one command. Here is how it works and how to use it.
Bottled Water Is Mostly Markup: The Product Is Often Tap Water, and the Margin Is the Whole Business
Aquafina and Dasani say on their own pages that they bottle municipal tap water. At Aurora's 2026 rate, a 20-ounce Dasani sells for roughly 2,800 times the water inside it.
BookBub: The Author Platform With 15 Million Power Readers
BookBub's 15 million subscribers are the most purchase-ready book audience on the internet. Here's how to set up a free author profile, understand featured deals, build followers, and use BookBub's ad platform for targeted book marketing.
The Can Is the Product: Ball Corporation and the Toll on Almost Every Aluminum Drink
Ball Corporation stopped making the mason jar it is famous for and became the company that makes the can. It ships about 103 billion of them a year and gets paid whatever is inside.
Autonomous AI Pentesting: How Strix Finds Real Vulnerabilities in Your App
Strix is an open-source tool that runs teams of AI hacker agents against your app, finds vulnerabilities, and proves them with working exploits instead of false positives. Here is how it works, how to use it, and how to use it safely.
Audit and Fix Your Website's Craft with Emil Kowalski's Skills Repo
A practical, copy-paste guide to auditing any website against Emil Kowalski's design-engineering standards and applying the fixes: press feedback, real easing, hover gating, reduced motion, text-resize, and installable PWAs.
The Armed Forces Retirement Home: The Benefit Only the Toll Booth Owner Can Build
Fifty cents a month from every enlisted paycheck since the 1850s, plus the fines of the court-martialed, funds a retirement home with care guaranteed regardless of ability to pay. No private employer could build it.
Saltchuk: The $5 Billion Family Company Most People Have Never Heard Of, and Where Its Money Really Comes From
Saltchuk is a private Seattle shipping and energy empire owned by three sisters. It is not an 8(a) firm. Its money comes from the Jones Act, Navy contracts, and a Hawaii monopoly.
How to Use Polsia, the AI That Says It Runs a Company While You Sleep
A working guide to Polsia: what it actually does, how to drive it, what it costs, who built it and why, and which of its headline numbers hold up to a source check.
How Title Insurers Get Paid: A Nickel in Claims, Seventy Cents to the Agent, and Nobody Shopping on Price
Only about a nickel of every title-premium dollar ever pays a claim, and roughly 70 cents goes to the agent. The buyer who pays the premium almost never picks the insurer.
How Ticketing Companies Get Paid: The Fees Are the Toll, and It Collects Them for Itself
Live Nation's ticketing arm is about 12% of revenue but earns a ~37% margin while its concert business runs near 3%. The fees are the toll, and it collects them partly for itself.
How Taxi Medallion Owners Get Paid: A $1.3 Million License, Then the City Erased the Scarcity
A New York taxi medallion is a city-capped license to run a cab. It traded up to about $1.3 million, then the city let in uncapped Uber and the toll collapsed.
How Saltchuk Was Built: A Lawyer, a Money-Losing Alaska Shipping Line, and Forty Years of Never Selling
Saltchuk began in 1982 when a Seattle lawyer and a handful of partners bought a failing Alaska shipping line named TOTE. Here is how that became a $5 billion family empire.
How ROV Pilots and Saturation Divers Get Paid: A Rate Sheet That Meters Labor by the Vertical Foot
Union diving contracts meter pay by the vertical foot, and saturation divers earn a bonus that ticks over every hour, around the clock, sealed in a pressurized chamber for up to 28 days.
How Pro Athletes Get Paid: The "Salary" Is a Revenue Share, Capped by Formula and Gated by a Clock
Players split roughly half of league revenue, then a cap rations it. The 2025 NFL cap of $279.2M per club is an output of that formula, not a wage budget.
How Police Officers Get Paid: The Badge the City Buys, Then Rents Out at a Public Rate
Cities pay for the badge, then rent it to stadiums and film sets at a public rate the officer keeps, and federal law caps police overtime lower than firefighters'.
How Pharmacy Benefit Managers Get Paid: The Toll That Grows Fatter as the Drug Gets More Expensive
PBMs can bill a plan $100 for a pill, pay the pharmacy $20, and keep the $80 spread. The FTC found the Big 3 marking up specialty generics by hundreds and thousands of percent.
How Performance Rights Organizations Get Paid: A 1941 Antitrust Settlement and a Per-Song Hammer
Play music in your bar, gym, or store and you owe ASCAP and BMI. You can't refuse because of a 1941 consent decree, and the 'or else' is $750 to $150,000 per song.
How Oil and Gas Field Workers Get Paid: A Wage Built on the Oil Price, the Hitch, and a Day Rate
A roustabout's $46,960 median barely describes oilfield pay. It runs on 84-hour hitches, day rates, per diem, and a headcount that fell about 43% when crude crashed.
How Registered Nurses Get Paid: A Safety Rule Sets the Wage, and Becoming a Temp of Your Own Job Is the Fastest Raise
A California patient-safety rule caps patients per nurse and behaves like a hidden wage floor. And staffing-agency margins jumped from 15% to 62% during COVID.
How Mortgage Brokers Get Paid: A Sales Commission Federal Law Forbids From Following the Price
Federal law freezes the mortgage broker's fee so it cannot move with the interest rate. Before 2011 a higher rate paid more: $1,000 at 5 percent, $3,000 at 6 percent.
How Mortgage Bond Sellers Get Paid: Three Tolls on One Mortgage, Before It Ever Reaches a Bank
A single mortgage is tolled three times on its way to a bank: gain on sale at the lender, a guaranty fee at the agency, and a dealer spread on the bond desk.
How Coal and Metal Miners Get Paid: A Modest Wage, a Daily Rate Sheet, and a Tax on the Disease the Job Causes
Union coal miners are paid portal-to-portal on a daily rate sheet, and a $1.10-per-ton federal tax pre-funds the black lung disease their own work causes.
How LPNs Get Paid: A Line in State Law Caps the Paycheck, and the Prestige Employer Pays the Least
The 2024 median LPN wage is $62,340, but a state scope-of-practice line caps it, and hospitals pay LPNs the least of any major setting at about $55,380.
How Longshoremen Get Paid: A $39 Base Wage, a $243,000 Average, and Four Devices That Explain the Gap
Full-time West Coast dockworkers averaged $243,000 in 2024 while the contract base wage topped out near $39 an hour. Hours guarantees, a cargo royalty, a closed register, and automation bans explain the rest.
How Lineworkers Get Paid: The Check Spikes When the Weather Is Worst
Lineworkers earn a $92,560 median, but the real money is storm work paid at double time, so a journeyman worth $74/hr straight is worth about $149/hr in a hurricane.
How Insurance Agents and Brokers Get Paid: A Toll on the Premium, Heaviest in Year One
An insurance agent's pay is a slice of your premium, crushed into year one: on term life the agent can take 60 to 80 percent of the first-year premium, and lifetime commissions total only 5 to 10 percent.
How Funeral Homes Get Paid: The Toll the Grieving Family Never Sees, Collected From a Life Insurer
The largest U.S. funeral operator booked $292.1M in 2025 not from families but from life insurers, for steering customers into prepaid contracts. And it earns a fatter margin on graves than funerals.
How Career Firefighters Get Paid: Federal Law Says a 53-Hour Week Is Straight Time
Federal overtime law exempts firefighters, so a 53-hour week is straight time, the shift is 24-on/48-off, and the hourly rate is spread across a 2,912-hour year.
How Financial Planners Get Paid: A 1% Toll on Money That Is Only Passing Through
The signature charge is a 1% annual fee on the client's whole balance, levied every year whether the market rises or falls. It is a toll on money in motion, not a wage.
How EMTs and Paramedics Get Paid: A Federal Rule Pays for the Ride, Not the Care, and the Wage Sits Near the Floor
Medicare pays for the ambulance ride, not the care given during it, so EMTs post a $41,340 median and paramedics $58,410 despite up to 1,800 training hours and a national exam.
How Elevator Mechanics Get Paid: A Six-Figure Trade Whose Door Opens Once Every Two Years
Elevator mechanics earned a $106,580 median in 2024, double the all-jobs figure, with no degree needed. The gate is an apprenticeship each local runs about once every two years.
How Domain Name Registries Get Paid: A $10.26 Toll on Every .com on Earth
Verisign collects $10.26 a year on every .com name under a government-sanctioned monopoly, pays ICANN just $0.2575, and books net margins near 50 percent.
How Credit Rating Agencies Get Paid: The Judged Hire the Judge, and Three Firms Collect the Toll
The bond issuer being graded pays the rating agency, the investors it protects pay nothing, and federal rules and bond indices force nearly every issuer to buy a rating from one of three firms.
How Court Reporters Get Paid: A Per-Page Toll on the Official Record
A court reporter earns a statutory per-page fee on every page of transcript on top of salary. Federal daily copy runs $7.30 a page, and every party pays the toll again.
How Correctional Officers Get Paid: The Post Cannot Go Empty, So the Overtime Is Mandatory
A prison post must be staffed every hour, so a short-handed shift becomes a forced double. One Connecticut officer booked $187,000 of overtime on a $71,000 base.
How Certificate-of-Need Incumbents Get Paid: A Permission Slip Your Competitor Has to Ask You For
In 35 CON states, an incumbent hospital is legally notified when a rival wants to build a bed or buy an MRI, and can drag the approval out about two years before any court fight. The FTC calls it a shield.
How Card Networks and Issuing Banks Get Paid: The 21-Cent Cap That Became a Floor
A big bank spent about 4.1 cents to authorize a 2023 debit swipe and collected 23 cents in interchange. The 21-cent federal cap became a floor, not a ceiling.
How Car Dealers Get Paid: A State Law Makes It Illegal to Sell You the Car Any Other Way
New-vehicle sales were 49% of AutoNation's 2025 revenue but only 13% of gross profit. The dealer isn't paid to sell you a car. It's paid because state law bans the maker from selling it directly.
How Bounty Hunters Get Paid: No Recovery, No Fee, and an Arrest Power From an 1872 Metaphor
Bail recovery agents earn 10 to 20 percent of a bond, but only if they physically produce the defendant, backed by an arrest power that traces to an 1872 Supreme Court metaphor.
How Bail Bondsmen Get Paid: A Non-Refundable 10 Percent Toll on the Door of the Cell
A licensed insurance agent unlocks a jail cell for a fee of 10 percent of bail, and keeps it whether you are acquitted or not. The industry paid under 1 percent in losses.
How Beer Distributors Get Paid: A State-Guaranteed Toll Booth You Can Inherit
Two state laws force every case of beer through a licensed wholesaler and make that wholesaler almost impossible to fire. A single mid-1970s distributorship is now a $31 billion fortune.
How Air Traffic Controllers Get Paid: The Job With a Legal Deadline to Start and a Legal Deadline to Quit
You generally must be hired before you turn 31 and you are forced out at 56. That legislated age window plus an academy that takes under 10% of applicants produces a $144,580 median for a job that needs an associate's degree.
How Actors and Film Crew Get Paid: The Health Insurance Belongs to No Job, and You Re-Earn It Every Six Months
Film crew must clear 600 covered hours to switch on health coverage, then 400 every six months to keep it. Performer scale runs about $1,246 a day plus residuals.
Never let a model pull the trigger: the action boundary for autonomous AI (Part 4 of 5)
The dangerous moment is when a model’s output causes an effect: a payment, an email, a deleted record, a deploy. Put a deterministic gate and a real kill switch between the model and the action.
How Union Tradesmen Get Paid: The Wage Is Only Two Thirds of It, and You Get Paid to Learn
Published wage sheets show pipefitters and operating engineers at packages from 76 to 131 dollars an hour, with a third of it going to pension, annuity and health.
How Railroad Workers Actually Get Paid: The Basic Day, the Mileage Rate, and a Pension System Nobody Talks About
Freight crews are paid by the mile, not the hour, off a basic day written in 1919. And they have a separate federal pension that pays more than double Social Security.
How Harbor Pilots Get Paid: The Job Where a Government Board Votes on How Many People May Have It
State boards cap the number of harbor pilots and set the price of their labor by formula. Published state figures put one port's pilots at 449,591 dollars each.
How Auto Workers Get Paid: The COLA That Was Won in 1948, Surrendered in 2009, and Bought Back in 2023
The UAW restored its cost of living allowance after fourteen years. The contract text shows what it actually pays, what a healthcare diversion takes back, and who still has no pension.
How Airline Pilots Actually Get Paid: What Is Written Into the Union Contracts
Pilots are not paid for the hours you would guess. Inside the credit hours, rigs, guarantees, snap-up clauses, and scope caps in the major US pilot contracts.
The Yellow Ribbon Reintegration Program: $181 Million to Reach the Half of the Force That Lives Nowhere Near a Base
DoD spent $181.329M on YRRP in FY2013, below the $207.8M projected, to link Guard and Reserve families to benefits. The honest gap is proof it works.
WHINSEC: A $10 Million School, a New Name, and a Closed Roster
WHINSEC, the Army school that replaced the School of the Americas, requests about $10.45M for FY2026, all Army O&M, and closed its student roster in 2016.
Unified Platform and JCC2: The Software Spine of U.S. Cyber Command
A June 2020 GAO review found Cyber Command's Unified Platform cost more than 5x its starting estimate, about $589 million projected, plus $224 million for JCC2.
The buried shells the Pentagon still cannot price
DOD's 2003 estimate to clear unexploded ordnance from 10,444 operational ranges ran from $16 billion to $165 billion, a tenfold spread GAO called unreliable.
The Expert's Trap: When Knowing Fifteen Things Is the Depreciating Asset
A defense data job wants fifteen disciplines at once. Here is why a decade of that expertise depreciates on three fronts at the same time, and what to bolt underneath it before it does.
The Space Test Program: the Pentagon's thirty-million-dollar ride to orbit
For about $30 million a year (FY2025 enacted: $30,279,000), the Space Test Program has flown 665 military experiments to orbit over 58 years, mostly as rideshare.
The sterile fly gambit: how the screwworm was eradicated, and why it is coming back
US ranchers gain roughly $900 million to $1 billion a year from screwworm eradication, a USDA program of sterile flies dropped from planes now facing a live re-invasion.
The $108 Million Classroom: The Six Pentagon Schools No Audit Can Price
DoD's six Regional Centers cost $108.8M enacted in FY2025 and reach 70,000+ foreign security officials, yet GAO once found no way to measure the payoff.
The Chemical Weapons Buried Under American Yards, and the Bill to Dig Them Up
The Army's buried-chemical-weapons cleanup carries a projected $2.5B to $17B 30-year cost, while Spring Valley DC alone ran to roughly $300M actually spent.
Operation Deep Freeze: The Military Airline That Keeps Antarctic Science Alive
Since 1955 the Air Force has airlifted NSF science to Antarctica. GAO called NSF's $275M McMurdo rebuild estimate unreliable, and one 1976 icebreaker holds the chain.
Nunn-Lugar: the $13 billion that pulled 7,600 Soviet warheads out of service
Nunn-Lugar deactivated more than 7,600 ex-Soviet nuclear warheads and scrapped 902 ICBMs for under $13 billion since 1992, Shchuchye cost overrun and all.
The Heat Ray That Never Fired: Inside the DoD Non-Lethal Weapons Program
GAO found the DoD Non-Lethal Weapons Program moved about $817M in budgeted funds over 12 years, spent at least $386M, and invented no new weapon of its own.
The Military Working Dog Program: The Mission Worked, the Kennels Did Not
A DoD Inspector General review found four military working dogs died in aging Air Force kennels from FY2021 to FY2023, inside a program costing up to $85,000 per dog.
The Homeowners Assistance Program: $555 Million to Cover a Loss the Government Caused
The Army-run Homeowners Assistance Program drew a $555 million stimulus appropriation in 2009 to cover home-sale losses for military families.
One Manager for Every Bomb: The Navy's EOD Technology and Training Program
The Navy single-manages every US military bomb-disposal tool on $43.8M spent in FY2024, yet GAO says DoD still cannot say what its whole EOD force costs.
The DoD Forensic Enterprise: The Army's Crime Lab That Went to War
DoD estimates it spent $800 million to $1 billion in war funds on expeditionary forensics from 2005 to 2012, per GAO, with weak cost visibility.
The $13 Million Prisoner: The Army's Job as Owner of the DoD Detainee Program
Guantanamo ran about $380 million, or $13 million per prisoner, in FY2018, while the Army executive-agent role that governs DoD detainee policy costs almost nothing.
DC3: The Pentagon's Volunteer Red Team, and the $300 Million Estimate Behind It
DC3 has processed 50,000+ vulnerability reports since 2016; its DIB program claims $300M saved, but that headline figure is a derived estimate, not audited money.
The Pentagon's Fingerprint Machine: DoD Biometrics and the Database That Knows Who You Are
DoD planned about $3.5 billion on biometrics for FY2007 to 2015 and built a 30-million-record identity database that helped capture 1,700, then lost Afghan data.
Anti-Tamper: The $50-Million Office That Guards the Secrets Inside America's Exported Weapons
The Air Force runs DoD's anti-tamper office on about $50 million a year (PE 0605024F); the widely cited $499M figure is a six-year contract ceiling, not money spent.
The Defense Production Act's Title III: A Few Billion Dollars, and Almost No Scorekeeper
A small Air Force office committed about $3.6 billion in DPA Title III factory investments from FY2018 to FY2024, and GAO found almost no one measured whether it worked.
The Defense Logistics Agency: the quiet backbone that feeds, treats, and fortifies the force
DLA obligated about $52.6 billion in FY2024 as the Pentagon's single Executive Agent for food, medical, and construction supply; GAO once found $7.1B in excess stock.
The Military's Language Schoolhouse: Monterey, Lackland, and a Skill the Market Won't Supply
The Defense Language Institute trains more than 2,500 resident students a year at Monterey in a dozen-plus languages, yet its annual dollar cost is hard to trace.
The Two-Million-Dollar Board That Governs Half a Million Tons of Explosives
The DoD Explosives Safety Board ran on a $1.92M FY2002 budget while governing ~493,000 tons of excess ammunition and a projected $3B disposal liability.
The Jammers That Silenced the Trigger: CREW, JIEDDO, and the Single Manager Who Never Quite Ran the Show
More than $18 billion went to JIEDDO through FY2011, funding the CREW jammers that beat roadside bombs, even as a single manager could not stop the duplication.
The Army's Counter-Drone Office: A Billion-Dollar Answer That Got Reorganized
Army counter-drone funding requests grew from $551.5 million to about $1.2 billion by 2025, per GAO, then the Pentagon moved to dissolve the office it built to run it.
The lab behind the MRE: cheap research, an expensive stockpile
GAO found DLA sat on a 5-million-case MRE war reserve that Army and Marine officials called 'probably not necessary.' The ration science is cheap; the stockpile isn't.
The Chemical and Biological Defense Program: One Army, Twenty-Six Organizations, $1.4 Billion a Year
The Army runs DoD's chemical and biological defense program, about $1.4 billion in the FY2026 request, and GAO has faulted its fragmentation for thirty years.
Money as a Weapon System: The $7.8 Billion Commanders Handed Out in Iraq and Afghanistan
Congress appropriated more than $7.8 billion for CERP so commanders could hand out cash in Iraq and Afghanistan, yet auditors could never prove the projects worked.
CATCH: The Pentagon's Quiet Database That Waits for a Serial Offender to Repeat
The Pentagon's CATCH database logged 2,999 anonymous sexual-assault victim submissions and found 155 suspect matches by September 2024, at a cost DoD never breaks out.
The Pentagon's blast-injury brain trust, and the research money that nearly vanished
DoD blast-injury TBI research funding fell from $175M (FY2024) to $40.5M (FY2026), over 75 percent, though its Army coordinating office has held the job since 2006.
How to Build the Skills a Defense Data Platform Job Asks For, and What It Pays
A Secret-level data engineer posting advertised $182,000 to $235,000 this week. Every skill on the job spec, Iceberg to lineage to GitOps, is free to practice at home.
The Army Wants Reactors on Its Bases Again. Project Pele Is the $300 Million Test.
A June 2022 contract worth about $300 million to build a single prototype anchors the Army's push to put nuclear microreactors on nine US bases by 2028.
Stablecoins Make Money by Keeping the Interest on Your Dollars. Here's Who Pockets It.
Tether cleared over $10 billion in 2025 with no full audit. Circle hands half its interest to Coinbase. Who really owns the major stablecoins, and how the float-interest machine pays them.
Serving a Browser LLM Is Easy. Securing It Is the Part the Demos Skip.
Adding a browser LLM means loosening your CSP, pulling code and model weights from third parties, and rendering untrusted output. Here is how to ship it without opening a hole in your site.
When an AI agent writes your code, a poisoned instruction is remote code execution (Part 3 of 5)
If an AI agent turns a task file into code and runs it, then a poisoned task file is remote code execution by proxy. Here is how to secure the build plane, including the agent-writes-tasks worm path.
GitHub's Top Repos of 2026 Are Mostly AI Agent Add-Ons. Here's What a Small Business Can Actually Do With Each One.
The repos topping GitHub in 2026 are mostly AI coding-agent tools, several past 100,000 stars in under a year. Here is what each one really costs and who can actually use it.
Claude Fable 5 Costs $50 a Million Tokens Out. Stop Using It as a Task Runner.
Fable 5 bills output at five times input, so the frontier model should plan and judge while cheaper Claude tiers do the work. The model-router setup, with real 2026 prices.
A 128GB Mini PC Runs the AI Model Your $2,000 Graphics Card Can't Load
A 24GB graphics card cannot hold a 70B model; a 128GB APU like DGX Spark or Strix Halo can, just slower. When a unified-memory APU beats a GPU for on-prem AI, and what runs on one.
The 2 Billion Dollar Cat Litter: How One Nuclear-Waste Drum Shut the Only US Repository
About 2 billion dollars: what one drum of nuclear waste cost after the wrong, wheat-based cat litter reacted underground and shut the only US repository for years.
WIN-T: the Army's mobile battlefield network, halted mid-fielding after billions
The Army spent over $6 billion on WIN-T, fielded its mobile Increment 2 to 14 brigades, then halted further buys in 2017 after testers called it fragile.
Who Pays to Put Out the Fire: The Cost Curve, the Red Card, and the Machine Behind Wildfire Suppression
Federal wildfire suppression first crossed a billion dollars in 2000 and hit a record 4.4 billion in 2021. In 2023 it cost 3.2 billion to fight fires on the fewest acres since 1998.
Who owns the wildfire fleet: the private air-tanker companies and their military and government roots
The federal government pays private firms to fly its firebombers. This traces who owns them, from a Navy SEAL turned senator to Blackstone, Bain, and family logging outfits.
Who Gets Paid to Fight the Fire: The Rented Air Force, the Retardant Contract, and Who Foots the Bill
The government owns almost none of the air tankers it flies. One company holds a $1.3 billion sole-source contract for the retardant. Here is where the suppression dollars actually go.
St. Louis radioactive waste: Coldwater Creek, West Lake, and the fire creeping toward Manhattan Project residue
St. Louis purified uranium for the first atomic bomb. Eighty years later the government is still digging out the waste, and the cleanup bill keeps rising.
VH-71: how the new Marine One cost more than Air Force One, then was cancelled
The VH-71 Kestrel spent close to $3 billion, breached Nunn-McCurdy, and was cancelled in 2009 with zero operational presidential helicopters ever delivered.
The VA's Oracle-Cerner Record System: 11,000 Lost Orders and a Program That Was Paused, Not Cancelled
More than 11,000 clinical orders vanished into a hidden queue at the VA's first Oracle-Cerner hospital, inside a paused, multibillion-dollar records overhaul.
US-VISIT: the border-exit system Congress ordered in 1996 and still has not fully delivered
GAO in 2007 found DHS spent about $1.3 billion and delivered basically one-half of US-VISIT: biometric entry works, comprehensive biometric exit does not.
Designed Too Small: The Uranium Processing Facility and the Roof That Cost Half a Billion Dollars
A new uranium facility was designed too small to fit its own equipment, so the government paid about 540 million dollars to raise the roof 13 feet. Here is the ledger.
TSAT: the military internet in the sky that was cancelled before it flew
Roughly $2.5 to $3 billion was spent developing the Air Force TSAT satellite network, and it was cancelled in 2009 with no satellite ever launched.
SPOT: the airport program that spent nearly a billion dollars to read travelers' behavior
TSA spent about $900 million since 2007 on behavior-detection officers, then GAO found no validated evidence the method beat chance at spotting threats.
The Superconducting Super Collider: $2 billion, 14 miles of tunnel, then cancelled
Congress spent about $2 billion and bored 14 miles of tunnel near Waxahachie, Texas, then cancelled the Super Collider in October 1993. Here is the record.
Sentinel ICBM: how an 81 percent cost jump tripped the Nunn-McCurdy limit and the program survived anyway
Sentinel's cost estimate grew about 81 percent, from roughly $95.8 billion to $140.9 billion, driven by ground silos and launch centers, not the missile. The Pentagon kept it.
SBInet: the billion-dollar virtual border fence that watched 53 miles and was cancelled
DHS spent nearly $1 billion on SBInet and got 53 miles of Arizona border coverage before cancelling it in January 2011. What the GAO record shows.
RECA and the atomic veterans: paying the people the government's own bomb tests irradiated
RECA has paid about $2.69 billion to some 41,900 claimants since 1990. A July 2025 law revived and expanded it, finally reaching Trinity downwinders.
The Raise Written Into the Contract: The Jobs That Get a COLA the Private Paycheck Never Will
Delta pilots won 34 percent over four years, East Coast dockworkers 62 percent over six, and public retirees a COLA indexed to inflation. The at-will W-2 raise is whatever a manager decides.
Public Money in the Leaks: What the Paradise Papers, WikiLeaks, and the Epstein Files Actually Show
Universities routed endowment money through Bermuda shell companies, one Guantanamo detainee costs about 13 million dollars a year, and Jeffrey Epstein moved millions through elite campuses. Here is the documented record.
Where the Public Money Goes: The Index of Programs, Lifelines, and Boondoggles
A cited index of how the federal government actually spends money, program by program, from rural lifelines to nuclear boondoggles, read for both the cost and the public good.
The Frozen Paycheck on Capitol Hill: Politician Salaries, Inflation, and the W-2 Worker
Congress can give itself an automatic inflation raise. It has declined every year since 2009, freezing its own pay at 174,000 dollars, down about a third in real terms.
PILT and Secure Rural Schools: Paying Counties for Land They Cannot Tax
Interior paid rural counties a record 733 million dollars in 2026 for federal land they cannot tax. Part of that record was a side effect of a sister program lapsing.
The Pentagon Has Never Passed a Full Audit: What That Actually Means
DoD failed a seventh straight audit in FY2024, a disclaimer of opinion meaning auditors could not verify the books, not proof that money is missing.
The Pension Backstop: A $62 Billion Surplus and a Taxpayer-Funded Rescue
One federal grant sent $35.8 billion to a single pension fund. Here is how the PBGC backstops 30 million private pensions, and why one program needed a bailout.
Operation Fast and Furious: the ATF gun-walking case, what it cost, and what the record actually proves
ATF let roughly 2,000 guns walk to trace them to cartel leaders, lost most, and two turned up where Agent Brian Terry was killed. The oversight record, dated and sourced.
The Trucks You Are Not Supposed to Notice: How the Government Moves Nuclear Weapons Across America
An unmarked federal fleet of armored trucks and armed agents has driven nuclear weapons and bomb-grade material more than 140 million miles with no release. It costs about 450 million a year.
The Nuclear Waste Fund: A $49 Billion Account for a Repository That Was Never Built
Ratepayers paid about 22 billion dollars into a fund for a permanent nuclear repository. It never opened. Taxpayers now separately pay billions in damages, and the bill keeps rising.
Trailblazer: the NSA modernization program that spent more than a billion dollars and delivered little
The NSA Inspector General called Trailblazer an expensive failure costing more than a billion dollars, and said it could not track how the money was spent.
The NSA's Data Centers: Public Construction Money, a Classified Operating Budget, and a Building That Could Not Turn On
A roughly $1.5B NSA data center in Utah suffered about 10 electrical arc-flash failures over 13 months. Here is what the record shows about the money.
The NRO's Secret Headquarters and the Billions Congress Says It Never Saw
A roughly $304M spy headquarters near Chantilly and $1B to $3.8B in unspent NRO funds, revealed in 1994 to 1996, became a case study in classified budget oversight.
Future Imagery Architecture: the $10 billion spy satellite that was never built
An estimated $10 billion went into the optical half of the NRO's Future Imagery Architecture, cancelled in 2005 with no satellite delivered. Here is the record.
NPOESS: the weather satellite program that doubled in cost and was dismantled
A program meant to save money by merging two polar weather satellite fleets saw its cost estimate roughly double to about $15 billion before the White House tore it up.
NMCI: when the Navy handed its shore computers to a contractor in one of the biggest IT deals ever
The Navy outsourced its shore IT to EDS in 2000 for up to $6.9 billion, restructured it to a 10-year $9.3 billion deal, and GAO found only 3 of 20 targets met.
NLOS-LS: the billion-dollar missile-in-a-box that was 92 percent finished and still cancelled
The Army's networked precision missile flunked its 2010 White Sands test at 61 percent reliability against an 85 percent bar, and was cancelled after more than a decade.
EnhancedView: What NGA Committed to Commercial Spy Imagery, and What It Bought
In August 2010 NGA committed up to $7.3 billion over ten years to two commercial imagery firms. A war drawdown, a funding cut, and a merger followed.
NECC: the command system meant to modernize GCCS that was cancelled with little fielded
The Net-Enabled Command Capability aimed to modernize military command and control, was cancelled around 2010, and its funding went back to the old system.
The National Flood Insurance Program: 22.5 Billion in the Red, and Why That Was the Plan
In 2017 Congress erased 16 billion dollars of NFIP debt. Eight years later the program owed the Treasury 22.5 billion again. Here is where that money goes and why.
Constellation: the roughly 9 billion dollar plan to return to the Moon, cancelled in 2010
NASA spent about 9 billion dollars on Constellation before it was cancelled in 2010. Ares I flew once. Orion and heavy-lift work carried forward into SLS.
MUOS: five satellites reached orbit before the radios that could fully use them
The Navy's roughly $7 billion MUOS constellation was fully on orbit by 2016, but GAO found its advanced capacity sat largely unused for years as terminals lagged.
The MOX Facility: How a Plant to Burn Bomb Plutonium Became a $7.6 Billion Hole in the Ground
Pitched at 1.6 billion dollars and three years, it consumed about 7.6 billion over eleven and never processed a gram before the government cancelled it in 2018.
Modern Integration Environments: CI/CD with Jenkins, Kubernetes, and the Glue Between
Kubernetes now runs in production for 82 percent of container users. Here is how it fits with Jenkins, GitHub Actions, and the four numbers that tell you whether your pipeline actually works.
MEADS: the air-defense system the Pentagon decided not to buy, then paid to finish
In 2011 the US decided not to field MEADS, then spent about 804 million dollars finishing it anyway. What the tri-national program cost and why.
The Littoral Combat Ship: warships retired before 15, with mission modules that never fully arrived
Per-ship cost more than doubled to about $500 million, the swappable mission packages largely failed, and the Navy began retiring hulls at 13, 11, and under 5.
Land Warrior: the Army cancelled its wearable soldier computer, then sent it to war anyway
The Army suspended Land Warrior funding in February 2007 for weight and cost, yet 229 lighter ensembles deployed to Iraq that May and the work seeded Nett Warrior.
The KC-46 Pegasus tanker: the fixed-price contract where Boeing, not the taxpayer, ate the overruns
Boeing has written off around 8 billion dollars in KC-46 tanker overruns because the contract capped the government's cost, an unusual outcome in defense buying.
JTRS: the software radio family, the flagship that was cancelled, and the standards that outlived it
The Army cancelled JTRS Ground Mobile Radio in 2011 after years and billions with zero radios fielded, while sibling radios and the SCA standard survived.
JLENS: the $2.7 billion radar blimp that broke its tether and drifted across Pennsylvania
The Army spent about $2.7 billion over 17 years on a cruise-missile radar aerostat that broke loose in 2015, dragged its cable across Pennsylvania, and cut power to thousands.
JEDI: the $10 Billion Pentagon Cloud Contract That Was Litigated to Death
DoD cancelled its $10B JEDI cloud contract in July 2021 with nothing fielded. The waste was time and litigation, not cash burned. Here is the record.
Why a Global Health Plan Costs $500 and a US Plan Costs $1,200: The Four Things the Price Tag Hides
The cheap international plan and the expensive US plan are not the same product. The global one usually excludes the US, screens out the sick, and pays lower prices everywhere else.
In-Q-Tel: How the CIA Built a Venture Fund, and Where the Money Goes
In-Q-Tel is a nonprofit that turns roughly $100M a year of intelligence money into startup stakes. Here is the money flow, the record, and the critique.
Hunters Point Shipyard: the radioactive cleanup its own contractor faked
The Navy paid Tetra Tech EC to prove a Cold War nuclear site was clean. Two supervisors went to prison, and the company paid the U.S. $97 million in 2025.
Hanford's vitrification plant: a $4 billion job that passed $17 billion and poured its first glass 25 years late
A plant contracted for $4.3 billion in 2000 and due in 2011 began treating waste in October 2025, while the tanks it must empty keep leaking toward the Columbia River.
GPS OCX: the ground system so late the new satellites flew without it
OCX, the Raytheon GPS ground control system, was accepted in July 2025 about 8 years late, then cancelled in April 2026 after roughly $6.27 billion spent.
Global Hawk Block 30: the drone the Air Force tried to cancel as costlier than a 1950s spy plane
In 2012 the Air Force tried to retire its RQ-4 Block 30 drone as pricier than the U-2. Congress refused, costs fell, and the U-2 became the target instead.
The GAO High-Risk List and Improper Payments: The Watchdog That Quantifies the Waste
Federal agencies estimate roughly 3 trillion dollars in improper payments since 2003. Here is the watchdog that counts them, and why 'improper' is not 'stolen.'
The Ford carrier: the most expensive warship ever built, and the elevators that would not work
USS Gerald R. Ford cost about $13.3 billion, the most expensive warship ever built, delivered in 2017 years before its catapults and elevators actually worked.
FOGBANK: The Bomb Material the Government Forgot How to Make
The US stopped producing a secret warhead material in 1989. When it needed more in the 2000s, it could not remember how. Relearning cost about 92 million dollars and a year.
Federally Qualified Health Centers: The Safety Net That Never Stops Almost Running Out of Money
Federally Qualified Health Centers served 32.4 million patients in 2024, about 1 in 11 Americans. Here is where the two money streams go, and why one keeps hitting a cliff.
The FBI case files that never went digital: Virtual Case File, Sentinel, and a decade of IT failure
The FBI spent about $170 million on Virtual Case File and deployed nothing, then took Sentinel years over budget before it finally shipped in 2012.
Blue Force Tracking: the network that cut friendly fire, and what it cost to keep
FBCB2 gave crews a live map of friendly and enemy positions and is credited with reducing fratricide in 2003. The bill was bandwidth, cost, and fragmentation.
The air-traffic-control computers the FAA mostly threw away: the Advanced Automation System
Of about $2.6 billion the FAA spent modernizing air-traffic-control computers, roughly $1.5 billion was written off as wasted. Here is the full record.
The Expeditionary Fighting Vehicle: roughly $3.3 billion spent on a Marine vehicle that never fielded
Roughly $3.3 billion went into the Marines' EFV over two decades before Gates cancelled it in 2011, with no vehicle ever fielded and a full program projected near $15 billion.
DIMHRS: the billion-dollar military HR and pay system that was never fielded
GAO tracked about $766 million spent on DIMHRS by March 2008 and roughly $1 billion through FY2009 for a joint military pay system cancelled in 2010.
BioWatch: the billion-dollar bioterror alarm the government could not prove would work
DHS spent well over $1 billion on BioWatch air samplers, yet GAO found it could not reliably establish the system would detect a real attack.
When the Bills Rise Faster Than the Raise: A 2026 Playbook for Insurance, Energy, and Health Costs
Auto insurance jumped 22 percent, home premiums 24 percent in three years, and 2026 marketplace payments 58 percent, while real pay rose about 1. Here is what actually cuts each bill.
Counting America: how census technology failed in 2010, then came in under budget in 2020
The 2010 census handheld computers failed and reverted to paper, adding billions. A decade later the 2020 count finished near $13.7B, under its $15.6B estimate.
The DEA, $4 Billion in Cash, and Seizures With No Charges: What the Inspector General Found
The DEA seized over $4 billion in cash from 2007 to 2016, and $3.2 billion was forfeited administratively with no charge filed and no judge involved.
The Zumwalt Destroyer: A Stealth Warship Whose Big Guns Have No Affordable Shells
The Navy built three Zumwalt destroyers for roughly $24.5 billion, then found each guided gun shell cost about $1 million, so it never bought the ammunition.
DCGS-A versus Palantir: what the courts actually decided about the Army's intelligence system
Two federal courts found the Army broke the commercial-items law when it tried to build DCGS-A Increment 2 single-source, not that Palantir was better.
Total Information Awareness: the DARPA data-mining program Congress defunded in 2003
Congress killed DARPA's Total Information Awareness in 2003 via the FY2004 defense act, Section 8131. The post-9/11 goal was shared; the mass-data method was not.
The Policy Analysis Market: the DARPA prediction market cancelled within a day
DARPA's Policy Analysis Market applied a legitimate forecasting method to an indefensible subject and was killed one day after two senators exposed it in July 2003.
The mRNA bet: how early DARPA money helped seed the platform behind the COVID vaccines
DARPA put roughly tens of millions into unproven mRNA years before COVID. Here is the honest ledger of shared credit, from Penn bench science to Warp Speed.
ARPANET to the internet: the DARPA research bet that seeded a trillion-dollar economy
A modest, sustained ARPA and NSF research bet produced the open TCP/IP internet, a foundation the BEA links to a $2.6 trillion U.S. digital economy in 2022.
HTV-2: the Mach 20 glider that failed both flights, and the hypersonic data it left behind
DARPA's HTV-2 flew twice, lost contact both times, and never completed its Mach 20 glide, yet returned rare hypersonic-flight and materials data.
The DARPA Grand Challenge: the prize nobody won that launched the self-driving car
In 2004 no robot car finished DARPA's desert race. Nineteen months later five did. A few million dollars in prizes helped seed a multibillion-dollar industry.
From CALO to Siri: the DARPA AI bet that became the voice in your phone
DARPA spent about $150 million on the PAL cognitive-AI program from 2003 to 2008. A spinout called Siri commercialized the research, and Apple shipped it in 2011.
The high-risk research bet: how DARPA and AFRL are built for a few huge wins and many failures
A roughly $4 billion a year DARPA seeded the internet and stealth while most bets failed. The portfolio logic behind both, and the AFRL contrast.
Critical Access Hospitals: The 101-Percent-of-Cost Bargain Holding Rural Medicine Together
Medicare paid Critical Access Hospitals about 12 billion dollars in 2023 at 101 percent of their own reported cost. Here is why that mechanic exists, and what it buys.
Command Post of the Future: the software soldiers loved, and the cost of being locked into it
CPOF is a rare DARPA-to-battlefield success that scaled to about 6,000 systems by 2009. The oversight story is proprietary lock-in and sustainment cost, not failure.
The RAH-66 Comanche: about $6.9 billion spent, two prototypes flown, program cancelled in 2004
The Army spent roughly $6.9 billion developing a stealth scout helicopter, built two flying prototypes, then cancelled it in 2004 and redirected the future money.
COLA vs the Merit Raise: The Inflation the Government Indexes For, and the Wage It Leaves Behind
Since 2021 Social Security's cost-of-living adjustment has compounded about 27 percent. The typical private raise, even at a 20-year high, came in under that, and a lucky 2 to 3 percent bump fell far behind.
Deepwater: the Coast Guard boats that were lengthened until they buckled
Eight Coast Guard patrol boats stretched from 110 to 123 feet buckled and cracked, all pulled from service by November 2006, roughly $100 million wasted.
Coalition networks: CENTRIXS, the Mission Partner Environment, and the sprawl the Pentagon keeps trying to untangle
GAO found DoD building coalition networks largely in isolation while CENTRIXS grew into 40-plus enclaves, and the $998M MPE-S contract was a ceiling, not spending.
How the CIA paid for its black sites: over $300 million that the public budget never showed
The SSCI study found the CIA's secret-prison program cost well over $300 million in non-personnel costs, all routed through classified covert-action accounts.
Delete Your Translation Bill: Chrome's On-Device Summarizer and Translator Are Already Stable
Not the flashy chatbot — the boring chores that quietly meter. Chrome's Summarizer, Translator, and Language Detector run on-device, are stable for the web since Chrome 138, and cost nothing per call.
Camp Lejeune: the poisoned base water, and the compensation that took a generation to arrive
Marine base drinking water was contaminated from 1953 to 1987. Federal science documents the harm; the remedy took decades, and fewer than 1% of claims are paid.
The C-27J Spartan: brand-new cargo planes flown almost straight to the boneyard
The Air Force spent about $567 million on roughly 21 C-27J airlifters, then retired the near-new fleet to the desert boneyard. Most were reused, not scrapped.
Military burn pits and the PACT Act: a toxic exposure the government denied, then paid for
For years the VA denied about 78% of burn-pit claims. The 2022 PACT Act replaced impossible proof with presumption, at a decade cost CBO scored near $277 billion.
Teleport Is Still Legally Named Gravitational. Every Rename Since Tells You Where The Money Moved.
Sign a Teleport contract in 2026 and you are contracting with Gravitational, Inc. The company has renamed itself three times, and each rename marks a real decision. Here is what changed and why.
The Black Lung Trust Fund: A Coal Tax That Never Quite Covers the Miners It Pays
The Black Lung Disability Trust Fund owes about 4.6 billion dollars to the Treasury, having borrowed almost every year since 1979, even as the disease it pays for resurges.
Automated and Manual Testing: Where Cypress and Selenium Fit, and Where a Human Still Wins
Google found about 1.5 percent of its test runs are flaky and nearly 16 percent of its tests flake. That is the number that decides how much end-to-end automation you should actually write.
IVAS: the Army's $22 billion AR goggle ceiling, and the $1.8 billion it actually spent
The Army's IVAS goggles carried a contract ceiling near $22 billion, but GAO found about $1.8 billion actually spent and roughly 10,000 headsets headed to storage.
Future Combat Systems: about 18 billion dollars spent on an Army program cancelled before it fielded its core
The Army spent roughly 18 billion dollars on Future Combat Systems before the 160 billion dollar program's manned-vehicle core was cancelled in 2009. What the record shows.
Amtrak's Long-Distance Trains: What It Costs to Keep Rural America on the Rails
The Sunset Limited cost about 578 dollars in federal subsidy per passenger, yet for hundreds of rural towns Amtrak's long-distance trains are the only intercity transit there is.
The Alaska Marine Highway: What It Costs to Run the Ferry That Is the Only Road
Alaska's ferries recover only about 31 percent of their operating cost from fares, yet they are the only road to Juneau, a state capital you cannot drive to.
The Airborne Laser: a 747 with a laser cannon, and roughly $5 billion to the boneyard
The Airborne Laser cost over $5 billion, shot down a boosting missile in 2010, and still ended up scrapped. Here is what the record shows, from GAO and MDA.
ECSS: the Air Force ERP that cost about $1.1 billion and delivered no fielded capability
By the Air Force and Senate PSI accounting, roughly $1.1 billion went to a logistics ERP cancelled in 2012 with no significant military capability.
Agent Orange: the Vietnam herbicide and the presumption system it forced into existence
The government sprayed 19 million gallons of herbicide over Vietnam. The compensation system took twenty years to build and is still adding conditions.
The X-51A Waverider: an Air Force scramjet that failed twice, then set a record
An AFRL scramjet demonstrator lost three of four vehicles across 2010 to 2013, then on May 1, 2013 burned for 3.5 minutes at Mach 5.1, a record.
DASNet, the Afghan Mission Network, and what the Taliban inherited
There is no documented Afghan network named DASNet. The verifiable record is the Afghan Mission Network and the biometric and pay systems left behind in 2021.
The Afghan Air Force G222s: about $486 million bought, $32,000 in scrap recovered
SIGAR found the US obligated about $486 million for 20 G222 transports for Afghanistan; 16 were cut up at Kabul and sold for roughly $32,000 in scrap.
The forest camouflage bought for an army in a country that is 2.1 percent forested
SIGAR estimated DoD spent up to 28 million dollars more than needed on Afghan army uniforms in a proprietary forest camouflage, out of roughly 94 million.
The 7 billion in equipment left in Afghanistan, and the 83 billion dollar myth
DOD reported about 7.1 billion in US-funded equipment remained in Afghanistan in 2021. The viral 83 billion figure is the 20-year force cost, not weapons seized.
The 43 million dollar gas station: a Pentagon task force that could not explain its own bill
SIGAR found a DoD task force spent about 43 million dollars on one Afghan gas station, roughly 140 times a comparable Pakistani one, and could not document why.
Aerial Common Sensor: the spy plane cancelled because its sensors were too heavy for the jet
The Army cancelled Aerial Common Sensor on January 12, 2006 with no aircraft delivered, paying a roughly $200 million termination fee after the payload outgrew the airframe.
The A-12 Avenger: the stealth jet cancelled in 1991 that took 23 years to litigate
Roughly $2 billion was paid to build a carrier stealth jet that never flew. Cheney cancelled it for default in 1991, and the fight lasted until 2014.
The 340B Drug Pricing Program: The Hundred-Billion-Dollar Discount That Costs Taxpayers Almost Nothing
340B moved roughly 100 billion dollars in discounted drug purchases in 2025 while costing taxpayers almost nothing. Here is how safety-net providers keep the spread.
WebGPU Shipped in Every Browser. You Can Finally Delete the Paid AI Fallback.
WebGPU is now on by default in Chrome, Edge, Firefox, and Safari. That is the one shift that turns on-device AI from a Chrome demo into a default you can build on. What changes, and the honest gaps.
Alaska Bypass Mail: The Freight Subsidy That Lives Inside the Postal System
USPS says it paid 118.9 million dollars for Alaska Bypass in 2024 and collected 38.2 million in postage. It never touches the freight. Another agency sets the price.
Simulink and Model-Based Design: There Is No Such Thing as a DO-178C Certified Code Generator
MathWorks states in writing that its DO Qualification Kit does not prequalify tools. Automotive is different: TÜV SÜD certifies the same tools for ISO 26262 ASIL A to D.
Render.com After the 1 August Plan Change: The 20x Bandwidth Cut and When It Is Actually the Right Host
Render's legacy workspace plans hit their auto-migration date on 1 August 2026. Included bandwidth fell 20x: Hobby 100 GB to 5 GB, Pro 500 GB to 25 GB. What it costs now.
Radiance Technologies: What 100 Percent Employee-Owned Looks Like in the Filings
Radiance's ESOP trust carried 191.3 million dollars of company stock at the end of 2024, about 98 percent of the plan's assets. Its plan effective date is 2008, not 1999.
The Vector Database You Were About to Buy Is a For Loop
A 22 MB embedding model runs in the tab and a few thousand chunks rank in a plain JS loop. On-device RAG needs no server, no vector database, and no per-query bill.
MATLAB: What It Actually Is, and Why Almost Every Article About Its Price Is Now Wrong
The perpetual Home license is gone. MathWorks now sells Home at 165 dollars a year, Student at 119, and the free MATLAB Online tier caps you at 20 hours a calendar month.
What the Web Actually Runs On: jsDelivr Says a Korean Font, Not React
A Korean webfont was jsDelivr's single biggest bandwidth consumer in June 2026 at 16.0% of the whole network. React ranked #535. What a CDN byte ranking really measures.
Jama Connect: The Trace Matrix Is Gone, and So Is 21 CFR 820.30
Jama Connect's coverage is just actual relationships divided by the ones an admin's rules expect. Configure no rules and the tool reports nothing wrong, forever.
What Hugging Face Actually Is: The Hub, the Libraries, and the $0.10 'Generous Free Tier'
Hugging Face markets a 'generous free tier'. Its pricing page says $0.10 a month, as of July 2026. Git LFS is gone, the CLI is now hf, and transformers is on v5.
How Jama, Cameo, MATLAB, Simulink and AFSIM Fit Together, and Where the Glue Is a Technology Preview
Jama's datasheet calls the Cameo connector seamless. Dassault's own docs for that connector call it a technology preview, sync one way, and ship a published bug list.
Cubic Corporation: The Board Took 75 Dollars a Share When 78 Was on the Table
Cubic's board took 75 dollars a share from Veritas over a live 78 dollar bid, leaving roughly 94 million behind. Five years on, the debt is eating the company.
Cameo Systems Modeler, CATIA Magic, and the MBSE Benefits Nobody Measured
Dassault renamed Cameo Systems Modeler to Magic Cyber-Systems Engineer by late 2020, and kept both names live. A review of 847 MBSE papers found 2 that measured anything.
Teleport's Top Offering Is Certificates That Expire By Default In 12 Hours. Here's What That Actually Buys You.
Teleport's pitch is that you shouldn't have secrets to steal, just certs that die in 12 hours. I read the 109-page analyst report behind their Leader banner. They're one of nine.
Prompt injection when the data is the attack (Part 2 of 5)
Direct prompt injection is the user typing the attack. Indirect injection hides it in a web page, a document, or a support ticket your model reads. Here is how to keep ingested content from becoming instructions.
The Agnelli Chain: Four Boxes, One Loyalty Share, and Control of Ferrari
John Elkann's indirect economic interest in Exor works out near 15 percent, yet the family holds 83.97 percent of Exor's votes. Here is the whole chain, box by box.
AFSIM: How the Air Force Ended Up Owning Boeing's Simulation Framework
Boeing spent about 35 million dollars of its own money on AFNES, then handed it to AFRL in 2013 with unlimited government rights including source code. It is now AFSIM.
When the Built-in LLM Won't Reach: Running Your Own Model in the Browsers Chrome Can't
Chrome's built-in LLM is Chrome-desktop-only. WebLLM and Transformers.js run full models on the visitor's own GPU in any WebGPU browser — Firefox and Safari included. How, and the tradeoffs.
Goodreads, LibraryThing, and StoryGraph: The Three Book Discovery Platforms You're Not Using
90 million readers on Goodreads, 2.5 million on LibraryThing, and a fast-growing audience on StoryGraph are discovering books without ever touching Google. Here's how to claim your author profiles, optimize your listings, and use each platform's discovery mechanics.
Your Browser Has a Free LLM Now: What Chrome's Prompt API Really Does to Your Per-Token Bill
Chrome ships an on-device LLM (Gemini Nano) a web page can call directly. No API bill, nothing uploaded. What's actually stable, what's still an origin trial, and how to build with it.
The three trust boundaries every AI system that acts has to draw (Part 1 of 5)
Most AI security advice worries about a chatbot’s replies. The real risk starts when agents build the system, it ingests attacker-influenced data, and its output spends money or ships code. Here is the mental model.
You Can Now Run Real AI in the Browser for Free, and Nothing Leaves the Visitor's Device
Google's LiteRT.js runs vision, image, and embedding models on your visitor's own GPU. No API bill, nothing leaves their device. What it does, and how to build with it.
Torch Technologies: The Employee-Owned Contractor That Voted to Stay That Way
Torch Technologies is owned entirely by an employee trust, and in 2023 its employee-owners voted by a 90 percent supermajority to become a public benefit corporation so no buyer could undo it.
Sierra Nevada Corporation: The Family-Owned Prime That Refused to Sell
Fatih and Eren Ozmen borrowed against their house in 1994 to buy a 20-person shop. Three decades later it won a 13 billion dollar Air Force contract, and they still have not sold.
Shield AI: The Autonomy Company Building Toward an IPO, Not a Sale
Shield AI reached a 12.7 billion dollar valuation with an AI pilot called Hivemind, and is raising its way toward a public listing rather than a sale, the venture path to staying independent.
The Boring Files on Your Site That Can Actually Bite You: Favicons, Logos, and SVG Uploads
A viral demo hid a tiny webpage inside a favicon. A favicon cannot run code by itself, but the real lessons for a site owner are worth your afternoon: image smuggling, SVG uploads, and the soft spot in your CSP.
Essential Air Service: What It Costs to Keep Small-Town America Flying
Essential Air Service subsidizes flights to 183 small and remote communities, sometimes over 1,900 dollars a passenger. Here is where the money goes, and why it survives every attempt to cut it.
Castellum: The Defense Roll-Up That Does Its Buying in Public
Castellum is a defense roll-up that buys small contractors with its own publicly traded stock. Its share count rose about 78 percent in two years, the price of doing private equity in public.
Arcfield: The Carve-Out Veritas Cut From Its Own Prime
Veritas Capital spent years assembling the contractor Peraton, then in 2021 carved a piece back out of it, named it Arcfield, and started the roll-up all over again.
Anduril: The Defense Company Built Not to Be Bought
A virtual-reality founder raised over 11 billion dollars and built Anduril into a 61 billion dollar defense company that buys its rivals instead of being bought.
When the Floor-Raisers Get Looted: The Fraud Inside the Programs Meant to Help
The same generous programs that raise people's floors get looted from the inside: rent-a-tribe lending, sham charities, sober-home Medicaid fraud, and a response that rarely gets the money back.
The W2 Trap and the Toll Economy: Why a Wage Is the One Input With No Guaranteed Return
Washington spends $7 trillion, borrows $1.8 trillion, and taxes wages hardest, while utilities, water, waste, and rail collect a guaranteed return on bills you cannot escape.
The Fraud the Justice Department Used to Chase
White-collar prosecutions are at a multi-decade low, the foreign-bribery law spent 2025 on pause, and the corruption unit went from 36 lawyers to 2. What the DOJ pulled back from, and kept.
Raise Your Floor: The Ladders That Trade the Jackpot for a Reliable Outcome
You can trade the jackpot for a high, reliable floor by climbing a structured ladder. The military benefit stack, medicine, the trades, and elective office, by the numbers.
Luck, Ladders, and Tolls: The Complete Series
A seven-part series, fully cited: how much of success is luck, the ladders that raise your floor, the legal edges and lotteries, the fraud inside the programs, and the toll economy.
Legal Edges and Lotteries: The Paths Gated by Who You Are
Some floors are raised by which category the law recognizes you in: set-asides, tribal contracting, senior water rights, the finance pedigree gate. Others are lotteries dressed as ladders.
Four Japanese Home Appliances Worth Importing to a US Home Office (and Which to Skip)
Japan quietly perfected some low-wattage home gear the US never adopted. Here are four worth knowing, with the honest voltage and cost reality check that the glossy roundups skip.
How Much of Success Is Luck? What a Forty-Year Simulation Reveals
Three physicists ran 1,000 simulated careers for forty years. The most successful people were almost never the most talented, just the luckiest. The evidence, fact-checked.
The Schemes You Are Not Supposed to Notice, and How to Stop Being Cannon Fodder
Most of these are not illegal. That is the point. Hidden fees, high-cost traps, and a tax code that taxes wages hardest. Here is how to pay less and move to the owner side.
Your AI Vendor Can Lock You Out Tomorrow. Here Is a One-Page Continuity Plan.
A real false-positive suspension cut about 60 employees off from Claude with a Google Form as the only way back. The fix is not panic, it is a simple AI continuity plan any small business can write in an afternoon.
Tribalco: From a Tribal 8(a) Startup to Two Billion in Federal Work
Kevin Moss and Andrew Stern launched Tribalco in 2004 through a Maine tribe's 8(a) program, graduated in 2013, and built a defense-telecom integrator with over two billion dollars in lifetime federal work.
The Tatitlek Corporation: When a Native Village Owns the Contractor
An Alaska Native village of about 90 people owns a government-contracting group that has drawn some 2.9 billion dollars in federal work. Inside the ANC ownership model and its 8(a) advantages.
PeopleTec: The Cleared Contractor That Sold to Its Own Employees
Two SAIC executives built PeopleTec into a 600-person missile-defense firm in Huntsville, then chose 100 percent employee ownership over selling to a private equity buyer.
Noetic Strategies: The Set-Aside Ladder, One Rung at a Time
A woman-owned Huntsville software and cyber shop stacked four small-business certifications, teamed with a bigger neighbor, and climbed toward the exit the rest of this series ends at.
Mobius Consulting: How a Small Firm Teamed Its Way to a $567 Million Win
Melaine Privitera left Booz Allen to start Mobius Consulting. The woman-owned HUBZone shop teamed with Parsons and took majority control of a 567-million-dollar missile-defense contract.
Maximus: How a Welfare-Reform Contractor Became a Public Giant
David Mastran started Maximus in 1975 with 12,000 dollars. Fifty years on it is a 5.4-billion-dollar public company running the safety net's phone lines and eligibility screens.
You Don't Need the Most Expensive AI. You Need a Workflow.
Microsoft's security system topped a hard public benchmark not with one giant model but with a pipeline of 100-plus small agents. The lesson a small business can actually use: architecture beats raw model size.
Government Contract Vehicles, Explained: The Hunting License Behind Every Deal
IDIQs, GWACs, GSA Schedules, and set-aside vehicles decoded. Why a seat on a contract vehicle, not any single award, is the thing that makes a government contractor worth buying.
BecTech: The Combat-Systems Shop That Stayed Independent for 29 Years
Rebecca Sutton left a ManTech senior-VP job in 1997 to build BecTech, a woman-owned Navy combat-systems and missile-defense engineering firm. Nearly three decades on, she still owns it.
AVUM: The Founder-Led Shop That Survived on the Right Contract Vehicles
Randall Mora founded AVUM in 1991 and still runs it. How a small, founder-led defense-software shop endures for thirty years by holding the right government contract vehicles.
Amentum: The $14 Billion Contractor That Went Public Without an IPO
Private equity bought AECOM's castoff services arm, bolted on DynCorp and PAE, then took Amentum public through a tax-free Reverse Morris Trust instead of an IPO.
York Space Systems: The Whole Machine, From a Billion-Dollar Buyout to a $4.75 Billion IPO
AE Industrial bought a majority of York Space Systems in 2022 at about 1.1 billion dollars, grew it into the Space Development Agency's largest satellite supplier, and took it public in 2026 at roughly 4.75 billion.
Why a National-Security Investor Sits in Boca Raton, Not the Beltway
AE Industrial runs a roughly nine-billion-dollar national-security portfolio from Boca Raton, Florida, not Washington. The GE roots, the no-income-tax draw, and the DC office it finally opened in 2024.
Wayne Garrett Co-Founded a Turbine Company and Kept Its Books. Both Earned Him a PE Seat.
He co-founded Power Systems Manufacturing as its CFO in 1998, rode it through two sales, and spent years as AeroEquity's own finance chief before becoming an operating partner.
Thomas Churbuck Built a Turbine-Parts Company, Sold It Twice, and Became an Operating Partner
He co-founded a turbine-parts startup in 1998, sold it to Calpine for about 43 million dollars, and its assets later fetched 242 million from Alstom. A founder's path to a PE seat.
How to Read an AI Subscription Before You Pay for the Big Tier
A class action over Claude's Max plans is a reminder, not a verdict: the headline multiplier on an AI subscription is not what governs what you get. The rate limits are. Here is how to read one before you upgrade.
Peter Cannito Tripled Polaris Alpha and Sold It. Now He Runs Redwire.
As CEO he tripled a defense roll-up in two years and sold it to Parsons, then a PE firm handed him a space platform that just bought a drone maker for 925 million dollars.
The Plumbing: The Funds, LPs, and Holding Companies That House These Deals
Behind every take-public is a stack of entities: a management company, a general partner, a fund limited partnership full of pension money, and a chain of holding LLCs. How the money is actually housed.
Paul Fulchino Grew Aviall 30 Percent a Year and Sold It to Boeing for 1.7 Billion Dollars
As chief executive he built Aviall into the world's largest independent aircraft-parts distributor, growing it about 30 percent a year, then sold it to Boeing in 2006.
Marc Duvall Ran a Five-Billion-Dollar Aerospace Division. Then a PE Firm Hired Him.
For about twelve years he ran Collins Aerostructures, a five-billion-dollar division making engine nacelles and thrust reversers, and more than doubled it. That scale is the credential.
Mano Nazar Ran a Nuclear Fleet. AeroEquity Wanted That Kind of Operator.
He was president and chief nuclear officer of NextEra's nuclear division, running eight reactors and leading a 4.5-billion-dollar expansion that added over 700 megawatts. That is the credential.
Les Daniels: The Financier Who Has Been AeroEquity's Co-Investor for Twenty Years
Not an operator or a general but a private equity founder: he co-owned a turbine-components maker with AeroEquity, sold it in 2012, and became the firm's longtime aerospace co-investor.
The Army's Top General Took a Private Equity Seat: Jim McConville
The 40th Chief of Staff of the U.S. Army, who led 1.2 million people and a 185-billion-dollar budget, joined a defense private equity firm about a month after he retired.
Gary Mercer Spent 36 Years Engineering GE's Jet Engines. Now He Advises a PE Portfolio.
He capped a 36-year GE career as vice president and general manager of engineering at GE Aviation, running the design and certification of its commercial and military engines.
One Playbook, Many Starting Hands: From a Citizen With No Experience to a Tribal-Owned Firm in Government Contracting
A citizen with no experience, a cleared non-veteran, a veteran, a service-disabled minority veteran, a woman owner, and a tribal-owned firm can all run the same build-sell-roll play. The rules, the combos, and the free tools.
Built to Be Bought: The Cleared-Contractor Playbook for Selling the Company and Keeping the Compounding
One founder grew a small Navy shop to 300 million dollars, sold it to ManTech, then joined the buyer's private equity firm. Inside the build, sell, and roll-the-equity pattern.
The Built to Be Bought Series: All Thirty-Five Pieces, in Reading Order
A reading guide to the whole series on how cleared government contractors get built, sold, and rolled: the playbook, the operating partners, the SEC filings, the fund structure, and the founder on-ramps.
Amazon KDP A+ Content: The Free Upgrade That Doubles Your Book Page Conversions
A+ Content (formerly Enhanced Brand Content) transforms your Amazon book listing from a wall of text into a visual, conversion-optimized sales page. Here's how to set it up for free, what modules to use, and the before/after data from six book listings.
Bill Boisture Ran Gulfstream, NetJets, and Hawker Beechcraft. Now He Advises the Buyers.
A fighter pilot turned aviation chief executive who ran Gulfstream and NetJets and steered Hawker Beechcraft through bankruptcy to a 1.4-billion-dollar sale to Textron.
From the CIA's Cyber Center to a Private Equity Seat: Andy Boyd
He ran the CIA's Center for Cyber Intelligence. Two years after retiring, a defense private equity firm made him an operating partner and put him in charge of its cyber company.
How a Defense PE Firm Gets Paid: The Equity Stakes Written Into Its SEC Filings
AE Industrial owned 77.5 percent of BigBear.ai and, at the peak, 71 percent of Redwire. The public filings show how the firm and its operating partners turn portfolio stakes into money.
The UN's Money: Its Pension Fund, Its Staff's Taxes, and the Gift of Its Land
The United Nations runs a ninety-billion-dollar pension fund, operates a private tax system that keeps its payroll inside its own walls, and sits on Manhattan land a single family bought and gave it.
The Biggest Family Offices and What They Actually Hold
The richest American fortunes barely appear in the quarterly 13F filings everyone reads. Here is what the Gates, Dell, Iconiq, Walton, Bloomberg, and Bezos vehicles actually disclose, and why most of the money stays invisible.
What Taylor Swift and Travis Kelce Actually Own
Taylor Swift owns her master recordings outright while Travis Kelce turns fame into private-company stakes and a single public holding. Here is what the documented record shows the couple actually owns.
George Soros: The Fund, the Family Office, and the Foundation
George Soros's fortune runs through three ordinary structures: a hedge fund, a family office, and one of the largest foundations in the world. A look at the money and the machinery, from the 1992 pound trade to a spring 2026 SEC filing.
Sakana AI's Fugu: A Router Built to Survive a Model Being Switched Off
Ten days into a government-ordered shutdown of two Anthropic models, Sakana AI shipped Fugu, a router that treats every underlying model as swappable behind one endpoint.
Relais & Châteaux: The Family Hotels and the Founding of a Luxury Road
How two music-hall entertainers, a country inn on the Rhône, and one French highway grew into Relais & Châteaux, a family-run network of 580 independent luxury hotels and restaurants.
The Log-Periodic Power Law: A Physicist's Bubble Model, Applied to the S&P 500 and Nasdaq 100
The log-periodic power law is a physicist's model that says a speculative bubble leaves a readable signature in the price itself. Here is how it fits the 1987 S&P 500 and the 2000 Nasdaq, and why its power to actually predict a crash stays contested.
The 12 Federal Reserve Banks: How They Are Funded, Who They Hire, What They Pay
The twelve regional Reserve Banks are privately incorporated, fund themselves out of interest income, and by law send their profit to the Treasury. Since 2022 they have sent very little, and their presidents out-earn the Fed Chair. Here is how the money and the payroll actually work.
The Crypto Wars: When the US Tried to Control Encryption, and Lost
In the 1990s the United States treated strong encryption as a weapon and tried to control who could have it. Here is how key escrow, PGP, printed source code, and two federal courts ended that fight.
Aman: The Resort Brand for the Ultra-Wealthy, and Who Owns It
Aman built a luxury business on scarcity: tiny resorts priced five times the local market. Behind the brand sits a two-country court fight that left one man, Vladislav Doronin, in full control.
Before You Mandate AI on Your Team, Build the Guardrail, Not the Dashboard
A viral story about a junior developer fired over AI-written code, and a real Coinbase mandate, point at the same small-business trap: measuring AI usage instead of building the review step that catches confident, broken work.
Who Actually Borrows From Social Security
The Social Security trust fund is, by law, allowed to lend to exactly one borrower: the federal government that sponsors it, which then spends the money on everything else. That is both perfectly real and, to critics, the government owing itself.
To Act, Not to Seem: How Sweden's Wallenbergs Control an Empire Through Foundations
One family controls much of Swedish industry, ABB, Ericsson, Atlas Copco, SEB, Saab, not by owning it, but by controlling the foundations that own the holding company that owns it. Read the structure.
Shirtsleeves in Three Generations: How the Vanderbilts Lost the Greatest Fortune in America
Cornelius Vanderbilt died the richest man in America. Within two generations the family had no perpetual trust, spent the principal on mansions later demolished, and by a 1973 reunion, reportedly no millionaire among them.
Three Empires, No Owners: How Foundations and Holding Companies Rule Tata, IKEA, and Agnelli
India's Tata, Sweden's IKEA, and Italy's Agnelli family each solved the same problem the same way: put an entity at the top that no heir can cash out and no rival can buy, holding a minority of the shares but a majority of the votes.
Learn to Secure AI, Then Build With It: SANS Training and a BrainyCanvas Review
Two different AI problems for a business: how to adopt it without getting burned, and what to actually build with it. SANS answers the first with serious security training. BrainyCanvas is one cheap tool for the second. Here's who each is really for.
The Firm Behind the Show: Ruthless, Cavalry, and How a Political Podcast's Money Works
The team behind the conservative podcast Ruthless runs a private consulting firm and a private media company. Campaign-finance law opens a rare window on the dollars flowing through, and closes on almost everything else.
The Archetype: How the Rockefellers Made a Fortune Outlive Its Family
John D. Rockefeller was the richest American in history. His descendants preserved the fortune by turning it into institutions: a family office, generation-spanning trusts, and foundations. Read how.
A Thousand Trusts, and Then a Breakup: The Pritzker Family
The Pritzkers built Hyatt and Marmon and preserved the fortune through possibly the most elaborate trust structure any American family assembled. Then, uniquely, that structure fractured in a famous breakup.
The Dynasty That Became a Museum and a Bank: The Mellons
Andrew Mellon was, at his peak, the third-richest force in America and its Treasury Secretary. The family preserved the fortune through trusts and foundations, and today the name lives on institutions.
Who Controls the News: The Families That Own What America Reads and Watches
The New York Times, Fox, Hearst, and Condé Nast are controlled by families who own a minority of the equity, using the same two tools the wealthy use to keep any asset: a trust and a dual-class share.
A Candy Kitchen to $129 Billion: The Mars Family's Century of Silence
The family behind M&M's, Snickers, and Royal Canin is the third richest in America, worth $129 billion. They have never gone public, forbid photographs, and have kept the whole empire out of view for a century.
Two Hotel Dynasties: The Marriotts Who Kept It and the Hiltons Who Gave It Away
Two families built the American hotel. One kept the company, the control, and the money in the family. The other routed almost the entire fortune into a foundation and lost the company to private equity.
The Biggest Company You Cannot Buy Stock In: The Koch Family Empire
Koch is the second-largest private company in America at $125 billion in revenue, and you cannot own a share of it by design. The family built a private industrial empire and a political network to match.
The Legal and the Contested: How Canadian Fortunes Went Offshore
The Paradise Papers showed New Brunswick's Irving family using a Bermuda insurer to move profits out of Canada tax-free, a legal arrangement. A separate KPMG Isle of Man scheme was something the tax authority called a sham. The line between them is the whole story.
One Wildcatter, Three Families, and a Silver Gamble: The Hunt Oil Dynasty
H.L. Hunt turned a gambler's stake into one of America's largest oil fortunes, then split it across three families. The branches produced the Kansas City Chiefs, a quiet energy empire, and a market catastrophe.
Oil, a Trust, and the World's Richest Museum: The Gettys
A Depression-era trust turned J. Paul Getty's oil stock into a fortune that still pays his heirs, and a separate trust turned his share into the richest art museum on earth, all shadowed by a famous kidnapping.
Two Percent of the Company, Forty Percent of the Vote: The Ford Family's Century of Control
The Ford family owns under 2 percent of Ford Motor Company yet controls 40 percent of the vote, a lock held since 1956 that no outsider can buy. How a supervoting share keeps a public company in a family.
The Federal Reserve: The Fair Criticism, and the Myths That Bury It
The Fed has real, documented problems: a revolving door, huge Wall Street speaking fees, a 2021 trading scandal. It is also buried under myths. Here is how to tell the fair critique from the fiction.
The Machinery of the Family Office: What the Very Rich Have That You Do Not
A family office is a private investment, tax, and estate department for a single fortune. It captures advantages ordinary individuals cannot easily reach, including a tax deduction the rest of the country lost, and it only makes sense above a high threshold.
The Biggest Company You Have Never Heard Of: Cargill and the Family That Hides It
Cargill is the largest private company in America, moving $154 billion of the world's food. The family that owns 88 percent of it has more billionaires than any other and works hard to stay invisible.
The Protected Franchise: How State Law Made the Car Dealership a Heritable Family Fortune
In all fifty states, the law requires new cars to be sold through independent franchised dealers and bars automakers from selling to you directly. It is a protection almost no other business enjoys, and it makes each dealership a heritable asset.
The Hundred-Million-Dollar Gift: Bezos's Courage and Civility Award, and Where the Money Went
Jeff Bezos handed Van Jones and Jose Andres $100 million each, not $10 million, to give away to charities of their choosing. Some of it is well documented. A great deal of it is not.
The Man Who Refused to Found a Dynasty: Andrew Carnegie's Great Giveaway
Andrew Carnegie was once the richest man in the world. He believed that to die rich was to die disgraced, and he gave away nearly all of it, some $350 million, on purpose. The counter-case.
Big Bill Zeckendorf: The Man Who Played Monopoly With Manhattan, and the Grandsons Who Learned From His Fall
He assembled the land under the United Nations, bought the Chrysler Building, and built an empire on borrowed money, then lost it all in 1965. Two generations later his grandsons rebuilt the name by refusing debt.
Where the Ledger Balances: The States Running the Other Way
North Dakota carries a $63,300 taxpayer surplus while $1.35 trillion in pension holes sit elsewhere. Where American ledgers, schools, water, and premiums balance, and why.
How the Waltons Keep Half a Trillion Dollars: A Holding Company, the 990s, and a Bank
The richest family in America holds its Walmart stock in one entity, gave it away before it appreciated, moves it through charitable trusts, and owns a $28 billion bank. Read the structure from the filings.
One Night at a Five-Star Hotel Now Costs More Than a Week of Average W-2 Pay
A 2026 night at Hotel Bel-Air or Hotel del Coronado runs $1,300 to $1,900, more than a week of the average American's gross W-2 pay. Wages roughly doubled since 2000. Luxury, tuition, and the stock market did not wait around.
The Nonprofits That Do Not Look Like It: Golf, the Grid, and a Billion-Dollar Benefit Trust
A golf tour paying Tiger Woods $18 million, the operator of the Texas power grid, a workers-comp insurer, a credit union, and airline benefit trusts. All tax-exempt, all invested in the same markets, read from their filings.
Same Law, Five Outcomes: Nine University Endowments Read From Their Filings
The endowment tax signed a year ago today lands on nine universities in five different places, from Stanford's 8% to Caltech's exemption. Read from their tax returns, they all invest in the same markets and are all cutting at once.
The Fund That Is Not Allowed to Invest: Social Security Against Norway and Saudi Arabia
Social Security holds $2.56 trillion and earned 2.6 percent on it in 2025. Norway holds $2.2 trillion and owns 1.5 percent of every listed company on earth. Same idea, opposite mandate. Here is why, and who benefits.
Sovereign Money: Native American Banks and the Financial Edge of Tribal Sovereignty
Tribal nations are sovereign governments with real tax and financing powers, a growing Native-owned banking sector, and a $44 billion gaming economy. The advantage is genuine, the limits are real, read the record.
The Trademark Is the Business: Six Trade Associations Read From Their Filings
The AMA earns most of its money copyrighting billing codes, NAR owns the word Realtor, and Blue Cross licenses a color. Read from their tax returns, these nonprofits run on trademarks, and several are now under antitrust fire.
The Best Retirement Plan in America Belongs to the Government
The Thrift Savings Plan holds over a trillion dollars for seven million federal workers at a few cents per thousand. It is the quiet advantage of the wealthy, index-fund compounding, handed to ordinary people.
This Week's Commons: Heat, Water, and the Grid, July 2026
One July week: a 78-degree thermostat ask beside a Madison Square Garden wedding, western towns rationing incumbents while permitting growth, and a grid mandate upheld ahead of the grid.
The Working Ledgers: How Seven Research Institutions Invest, Stay Solvent, and Keep the Work Going
Read from their own tax returns, seven American research institutions run on completely different money, from MITRE's no-reserve pass-through to HHMI's $27 billion endowment. All of them compete in the markets you rely on too.
The Standing Institutions: Twelve Nonprofits You Know, and the Markets They All Depend On
The Smithsonian, Georgetown, AARP, the Kennedy Center, a pilots' union, and a university's own insurance company, read from their tax returns. Different missions, the same five ways to pay for it, and the same markets underneath.
The New Fortunes: How America's Biggest New Foundations Invest, Give, and Outgrow the Rules
Read from their tax returns: two dozen foundations holding well over $100 billion, where the chief investment officer often out-earns the president and the Nvidia boom is inflating corpuses faster than the law can make them give it away.
The Rules Beneath the Endowment: How 1969 Reshaped Every Foundation, and How 2025 Is Doing It Again
A 1969 law forced El Pomar to sell the Broadmoor, made every private foundation pay out 5 percent, and taxed its investment income. The 2017 and 2025 endowment taxes are its direct descendants. Read from the statutes.
Stanford's $52 Billion Balance Sheet: The Richest University Meets the Top Tax Bracket
Stanford holds a $40.8 billion endowment and $52 billion in net assets, invests it like a private-equity firm, and in 2025 landed in the top 8% endowment-tax bracket and cut 363 jobs. Read from the filing.
Neither a University Nor a Company: How Southwest Research Institute Funds Juno, Lucy, and New Horizons
The same San Antonio nonprofit runs engine test stands for fuel companies and leads NASA missions to Jupiter and Pluto. The 990 shows a client-funded lab that keeps a conservative reserve and reinvests the rest.
The Stranger's Gift: James Smithson, the Red Castle, and How a Bequest and the Government Built the Smithsonian
An Englishman who never set foot in America left his fortune to found the Smithsonian. His $508,318 in gold and 178 years of federal money co-mingled into a $6.5 billion hybrid. Read from the audited filings.
The $10 Billion Endowment That Gives Away Medicine: How Shriners Children's Pays for Free Care
Shriners runs pediatric hospitals and gives the care away, funded not by billing but by a $10.7 billion endowment. The filing shows a hospital system built like a foundation, and the market crash that changed it.
The Runway and the Ruler: Your Paycheck's Two Quiet Problems
Social Security's 2032 runway per the 2026 Trustees Report, a max earner's $447,000 against a $4 million counterfactual, the CPI's rebuilt ruler, and the insurance float that ends in Bermuda.
What the Rothschilds Actually Did With Money, and What the Myths Get Wrong
The Rothschilds did not use whole life insurance to rule the world. That is a sales pitch wrapped in an old conspiracy. Here is the documented record, and the real, legal wealth-transfer tools anyone can use.
The Quiet Shelf, Volume Two: A Century of Momentum, From Penrose's Orchard to a Powerball Ticket
Twenty more sets of books, read by age: a 1925 community trust turning one hundred, the Albertson stock gift that multiplied a foundation fourteen times, two Walton grandsons, and a billion that landed last year.
The Quiet Shelf: Nine Sets of Books That Show How American Giving Actually Works
Eight Rocky Mountain foundations and one $31 billion donor-advised machine, read from their own filings: a rubber fortune funding journalism, four Coors women around a table, and a $21 million fund stocking a school pantry.
Priced, Pooled, or Plundered: Three Ways to Run a Commons
Australia priced its water commons, Macquarie-era Thames Water extracted one, and the difference is a working playbook for any household that depends on either kind.
The $760 Million Nonprofit That Pays Tiger Woods: The PGA Tour's Tax Status Meets the Saudi Fund
The PGA Tour is a tax-exempt business league with $760 million in revenue and $3.9 billion in assets that paid Tiger Woods $18 million. Now the Saudi fund and the Senate are both circling. Read from the filing.
Who Holds a Public University's Money: The Oregon Foundations Behind OHSU and Oregon State
A public university cannot easily hold its own endowment, so a separate nonprofit does it. Read from their filings: two billion-dollar Oregon foundations, a $2 billion Knight gift, and a tax form worth double the audit.
The Margin Is Thin, the Portfolio Is Not: How Nonprofit Hospitals Actually Make Money
Nonprofit hospitals run on razor-thin operating margins, but their bottom line is decided by their investment portfolios. Scripps earned $206 million from operations and $958 million overall. Read from the filings.
The Rulebook That Set the Price of Every Home: The Realtors' Nonprofit and Its $418 Million Reckoning
The largest trade association in America owns the word Realtor, spent $86 million lobbying in one year, holds a billion in assets, and just paid $418 million to settle the case that dismantled its commission rule.
Two and a Half Billion Dollars, No Endowment: How MITRE Runs Six Federal Labs on Almost No Reserve
MITRE takes in $2.48 billion a year and holds essentially zero investments. The 990 shows a pass-through that turns federal money into a technical workforce and keeps only a few months of cushion.
The Candy Fortune That Raises Two Thousand Children and Controls a Public Company
A twice-bankrupt candy maker with no children gave his entire fortune to a school for the poor. A century later it is a $23 billion trust that educates 2,000 kids free and controls Hershey. Read the filings.
Sugar to Four Seasons: How Two Mauritian Families Diversified a Cash Crop Over a Century
On a small island, two Franco-Mauritian families turned colonial sugar into diversified empires spanning hotels, a bank, and textile plants in Bangladesh, as the guaranteed sugar price vanished. Read the record.
Half a Billion Dollars, Almost All of It Toys: How the Marine Toys for Tots Foundation Actually Works
It moved 25 million toys with eight people paid over $100,000 and a reserve invested in the same ETFs you can buy. The 990 shows a charity that spent into its rainy-day fund this year, on purpose.
The Genius Grants and the 5% Rule: How the MacArthur Foundation Decided to Spend Faster
A $9.3 billion endowment where the investment chief out-earns the president, famous for the no-strings genius grants, just broke philanthropy's most sacred number and pushed its payout past 7% to backfill federal cuts.
The Kennedy Money Machine, and How Everyone Since Has Copied It
The Kennedys did not own the Chrysler Building; they owned the Chicago Merchandise Mart, and it was a tax-advantaged cash machine. Here is what they actually did, and the legal versions anyone can use.
The Nickel a Day That Built Kaiser Permanente: From a Desert Hospital to a Hundred-Billion-Dollar System
Before Kaiser Permanente was a $128 billion system, it was a six-bed hospital in the Mojave and a nickel a day. How it started, who it was built for, and where it stands now.
The $116 Billion Nonprofit Whose Bottom Line Rides on the Market: Inside Kaiser Permanente
Kaiser is the largest nonprofit health system in the country, its operating margin is near zero, and its profit or loss is decided by a $54 billion investment portfolio. Its own CFO says it relies on returns to bridge the gap.
Save the Tax, Park the Savings: How Jackson Hole Became the Richest Place in America
Teton County, Wyoming has the highest per-capita income in America, and 77 percent of it is investment income. Wyoming's zero taxes, 1,000-year trusts, and anonymous LLCs are why. Read the statutes.
The Valuation Games: Seven Big Charities and What a Donated Good Is Worth
How a charity values a donated can of food or box of medicine decides how big it looks and how efficient it seems. Read from the filings: Feeding America at $1.97 a pound, a $3.7 billion drug foundation, and the fights in between.
The Endowment That Employs Its Scientists: The Howard Hughes Medical Institute, Read From Its $27 Billion Ledger
A recluse's tax dodge became a $27.6 billion machine that pays biologists to think. The 990 shows how it invests, what it owes, and why its highest-paid officer runs the money, not the science.
How Deathless Money Invests: Styles, Returns, and Risk Across America's Endowed Fortunes
Girard's trustees hold 50 percent equities, Hershey's deed locks 81 percent in one stock, small family funds sit in Treasuries, and a spend-down fund holds almost nothing. Seven investment styles, matched to the jobs they do.
The Hidden Owners: Who Really Owns the Hotel Jerome, and the Quiet Fortunes You Never Hear About
Aspen's most famous hotel is owned by a Houston billionaire who also owns the company that runs it. He is one of many enormous American fortunes held in private and on purpose. Read the record.
The Orphan's Dividend: What Girard's Money Actually Bought, 177 Years of Graduates Later
A WNBA Finals MVP, a two-time Pulitzer winner, Mobil's president, and the architect of Robert Morris's bank all grew up on one dead man's scholarship. The Girard endowment, measured in people.
The Museum That Is Also a $17 Billion Endowment: How the Getty Trust Pays for Free Art
The wealthiest art institution on earth charges nothing at the door because it lives on its investments. The filing shows an $8.4 billion endowment, a fire that nearly reached it, and the 5% rule that runs it.
The Codicil That Ran Two Hundred Years: Benjamin Franklin's Bet on Compound Interest
In 1789 Benjamin Franklin left 2,000 pounds to Boston and Philadelphia, not to spend but to lend and compound for 200 years. He projected 4 million. The trusts came due in 1991. Read what happened.
Let There Be Light, and a Reserve: Inside Fiat Lux, the University of California's Own Insurance Company
The University of California built its own insurance company, named it after its motto, and now holds $2.6 billion in it. Fiat Lux keeps the float commercial insurers used to pocket. Read from the audited filings.
The Biggest Charity in America Runs on Donated Food: Feeding America at $1.97 a Pound
Feeding America is the largest US charity by revenue, and about 90% of it is donated food valued at $1.97 a pound. The filing shows a national network, a single load-bearing number, and a 2025 hunger crisis.
The Federal Storehouse: Who Fills the Pool, Who Draws It, Who Rebuilds
Three states fund the pool, the average state draws $1.32 per dollar paid, and the flood program rebuilds the same house eighteen times. The federal storehouse, audited.
The Nonprofit That Runs the Texas Grid: ERCOT, the 2021 Freeze, and a Market That Clears Billions
A tax-exempt nonprofit keeps the lights on for 25 million Texans, funded by a fee on every megawatt-hour. Its filing shows $7.4 billion flowing through and a 2021 blackout that killed hundreds. Read from the record.
The Hundred-Billion-Dollar Reserve, and the Shell Companies That Hid It: Ensign Peak
A homegrown American church quietly built a fund rivaling the largest sovereign wealth funds, disclosed almost nothing, and was revealed only by a whistleblower and an SEC settlement. Read the record.
The Orchard at the Foot of the Mountain: Spencer and Julie Penrose, and Who Gets the Dividends Now
A Philadelphia black sheep sold a gold mine and kept the mill, built the Broadmoor outside a dry town, and left it all to an orchard-named trust. The 990 shows exactly how the money works 89 years later.
The du Pont Fortune, Two Centuries On: The Trusts That Still Pay, From Delaware to a Dome in Idaho
A fortune made in gunpowder in 1802 still supports its descendants and its home region, not as one pile but through trusts and foundations, while an eighth-generation heir skis off-grid in Idaho.
From the Apollo Guidance Computer to a $270 Million Reserve: How Draper Laboratory Sustains Itself
The lab that guided Apollo to the Moon is now an independent nonprofit that banked a cushion its FFRDC cousins never built. The 990 shows a contract shop investing $270 million like a small endowment.
The Biggest Charity in America Is a Wall Street Firm's Waiting Room
Fidelity Charitable takes in and gives out more money than any charity in America. It is the charitable arm of a family-controlled asset manager, the deduction is immediate, and the money can wait forever.
Why You Cannot Start a Bank, and Why the Rich Buy One Instead
America went from a hundred new banks a year to almost none. A charter is now a scarce, valuable asset, which is why the wealthy buy existing banks. Sam Walton's 1961 purchase is the template.
Two Quiet Dynasties: How the Crowns and the Kroenkes Hold Diversified Billions
A gravel company became a defense stake and a ski empire; shopping centers near Walmart became sports teams and the largest private landholding in America. Two families, one playbook, read from the record.
A Billion-Dollar Endowment on Three Legs: How the Broad Institute Pays for Itself
Eli Broad gave about $700 million and told the trustees to make it permanent. The 990 shows a biomedical powerhouse balanced on federal grants, philanthropy, and a $1.5 billion endowment, with a board of market titans.
Bill Daniels Repaid Debts He Didn't Owe. His Money Still Follows Orders.
A Golden Gloves boxer and carrier pilot brought television across 225 miles of Wyoming, repaid a bankrupt team's creditors $750,000 he didn't owe, and left a $1.7 billion fund still following his letter.
Other People's Money: How Insurance Float Built Berkshire Hathaway
Warren Buffett built a $176 billion war chest out of money that is not his: insurance float, the premiums held before claims are paid. How it works, and what it echoes for everyone else.
The Crown Jewel of Compounding: $120,000 to Warren Buffett in 1956, $1.24 Billion in 2024
A physicist handed the 25-year-old Warren Buffett $120,000 in 1956. In 2024 his daughter's trust moved Berkshire shares to the family foundation, and its assets went from $1 million to $1.24 billion.
Where the Money Lives: The Nature of Homes in Aspen and Ketchum
Two mining busts reborn as billionaire ski towns, where a $108 million house is held behind an LLC and left mostly empty, and the people who work there commute from an hour away. Read the record.
America's Hidden Sovereign Wealth: The State Funds Behind Alaska's Dividend and Texas's Schools
The US does run Norway-style sovereign wealth funds. They just sit in Juneau and Austin. Alaska cuts every resident a check; Texas oil land bankrolls its schools and universities. Read the numbers.
The Corporation That Certifies the Rockets: How The Aerospace Corporation Stays Solvent Without Building Anything
It builds no hardware, holds no endowment, and its biggest liability is a pension. The 990 shows an FFRDC that lives on Space Force contracts and competes in the market only through its retirement plan.
William Duer: The Founding Father Who Crashed an American Market First
Hamilton's first assistant treasury secretary borrowed, cornered, and defaulted his way into the Panic of 1792. Securities fell 20 percent in weeks. He died in debtors' prison.
William Bingham Was the Richest Man in Early America. You've Never Heard of Him.
Richest man in the young republic at 28, banker to the Revolution, two million acres of Maine, a Senate seat, and near-total obscurity. The fortune that survived is the lesson.
AI Stopped Being a Vendor Relationship and Became a Procurement Story
Western tech companies are quietly routing AI workloads to Chinese models like DeepSeek, Qwen, Kimi, and GLM. It isn't politics. It's unit economics, and the bill is doing the talking.
Thomas Willing Voted No on Independence. Then He Ran America's Money for 26 Years.
He voted against the Declaration twice and never signed it. Then he ran the first bank in America and the First Bank of the United States, and died at 89, reputedly the country's richest man.
Tench Coxe Drafted Hamilton's Masterpiece. His Land Paid Off a Century Late.
He wrote the first draft of the Report on Manufactures, preached cotton before the gin, survived the nickname Mr. Facing Bothways, and bought the coal land that enriched his descendants.
Stephen Girard Personally Kept the War of 1812 Afloat. Then He Gave It All Away.
A one-eyed French cabin boy became America's richest man, bought the dead First Bank's building, underwrote the 1813 war loan, and left nearly everything to orphans. The full ledger is complicated.
The Seven Ledger Laws in the AI Era: 1791's Playbook for 2026's Machine
AI adoption jumped from 55 to 78 percent in one year while inference costs fell 280-fold. The founders already ran this experiment, and their seven ledger laws say exactly how to position.
Samuel Meredith Held the New Nation's Purse for 12 Years. Then He Walked Away.
Washington made him the Constitution's first Treasurer in 1789. He served twelve years, resigned rather than serve Jefferson, and died on a wilderness estate in Wayne County.
Robert Morris: The Founding Father in the Capitol Dome Nobody Can Name
Five real Americans are named in the Capitol dome fresco. Four are household names. The fifth financed the Revolution, built the first federal offices, and died broke.
Oliver Wolcott Jr. Had the Worst Job in America: Following Alexander Hamilton
He climbed every Treasury rung to become the second Secretary, ran Hamilton's machine for six years in its creator's shadow, then reinvented himself as Connecticut's reforming governor.
Michael Hillegas Was America's First Treasurer. His Reward Came 103 Years Late.
He held the nation's purse for fourteen years while its paper money died at 500 to 1. The Treasury put his face on the ten-dollar gold certificate a century after he died.
James Wilson Drafted the Constitution and Died Hiding From His Creditors
He proposed the single presidency and the Electoral College, and sat on the first Supreme Court. He died at 55 in a colleague's house, dodging arrest for debt.
Henry Knox Hauled 60 Tons of Cannon Through a Winter. Maine Broke Him Anyway.
The bookseller who dragged Ticonderoga's guns 300 miles became the first Secretary of War, then bet everything on 576,000 acres of Maine. His widow inherited the debt.
Haym Salomon: The Immigrant Broker Who Kept the Revolution Solvent
Chicago put him in bronze holding George Washington's hand. The documented record: two British arrests, a reported death sentence, and an estate $560 in the red.
Gouverneur Morris Wrote 'We the People' and America Forgot the Rest of Him
The most quoted phrase in America was written by a one-legged New Yorker who stared down the Terror in Paris, drew Manhattan's street grid, and got left out of the pantheon.
George Clymer: The Founder Who Took the Jobs Nobody Wanted
Six men signed both the Declaration and the Constitution. The least famous one kept the army fed, volunteered for the worst assignments, and died a bank president, not a debtor.
1792 Is Trending: The Modern Headlines Hiding in the Founders' Ledgers
FTX reran Duer's 1792 playbook, the SVB backstop was Hamilton's rescue with bigger numbers, 2008 was the Panic of 1797 with mortgages, and the FTC's $12.5 billion fraud year is Scioto at scale.
The Ledger Lessons: 15 Founding Fortunes and What They Teach About Money in 2026
Fifteen men ran, financed, or crashed early America's money. Their endings sort into seven laws that map directly onto modern financial literacy, and the current numbers say we still fail the same ways.
The Common Course: Four Centuries of Collective Experiments, Read from the Ledger
From Plymouth's common storehouse to New Harmony, Debs, the kibbutz, Xiaogang village, and the DSA's New York, the collective experiments left four hundred years of data. The ledger grades them all the same way.
Albert Gallatin Ran the Treasury Longer Than Anyone. America Remembers the Man He Fought.
A Swiss immigrant ran the U.S. Treasury for nearly thirteen years, longest ever, cut the debt while buying Louisiana, then founded a university. Almost nobody can place his name.
The Quant Quake Stopped Being an Accident
Three 'worst since' headlines in eight months, the same crowded shorts every time. When a hedge fund blowup repeats on a schedule, it isn't bad luck anymore. It's how the market is wired.
Nation-State Malware, Protestware, and the Packages Your AI Pulls In
There is no official FSB package repository. The real threat is quieter: maintainer takeovers, protestware, and self-spreading worms in the dependencies your AI assistant installs. How to check if you are already compromised, stay clean, and keep your libraries current.
Build a Small-Business AI Front Desk: Setting Up an AI Agent and a Self-Sorting Inbox
A practical setup guide for two AI tools a small business can stand up this week: Connex AI's Athena agents for calls and chat, and Superhuman for an inbox that triages and drafts itself. Real steps, real examples, and honest costs.
Bluesky Custom Feeds: Become the Curator for 42 Million Users
Bluesky's AT Protocol lets anyone create custom algorithmic feeds. With Attie AI, you can build feeds using plain language — no coding required. Here's how to become the go-to curator in your niche.
Merchant Processors for Small Business: Keep More of Every Sale, and Don't Get Your Payment API Owned
Two ways small businesses lose money on card payments: the processor's cut and a broken payment API. A 2026 read on Helcim, Stripe, Stax, Dharma, surcharging, and ACH, plus the prompts and checks that keep your payments from getting owned.
Replace Mailchimp With a Claude Agent, a Free GitHub Cron, and Resend
Mailchimp bills per contact even in months you never hit send. A Claude agent writes the issue, a free GitHub Actions cron runs it, and Resend delivers it.
Why Use the Separate Anthropic API, and How It Actually Works
The Anthropic API is a separate product from your Claude subscription. Pay per token, build Claude into your own software, and reach features the apps do not have.
Which Terminal AI Coding Agent: Claude, ChatGPT, Gemini, and When to Run More Than One
There's no single best terminal AI coding agent. Claude Code, Codex, Gemini CLI, and Copilot each have a sharp peak and a real weakness. Here's which to run, and how many.
Claude on a Schedule and at Scale: Routines, GitHub Actions, the Batch Discount, and Model-Routed Gates
Put Claude on a schedule with routines, then run it cheaply at volume with GitHub Actions parallelism, the Batches API 50% discount, and model-routed gates.
Claude on the Desktop: What the App Does That the Browser Cannot
The Claude desktop app reaches your local files, runs local connectors, can drive your computer, and unifies Chat, Cowork, Code, and Dispatch in one window.
Claude Cowork: the Knowledge Worker's Version of Claude Code
Claude Cowork is Claude Code's counterpart for non-coders: an agent that operates your computer to do assembly work and hand back a finished deliverable.
Run /init First: How Claude Code Maps an Existing Project Before It Changes a Line
Claude Code's /init scans your project once and writes a CLAUDE.md it loads every session. Why that beats re-scanning each time, plus #, @, plan mode, thinking, and subagents.
Claude Code Without the Terminal: the Desktop App, the IDEs, the Cloud, Dispatch, and Building on the API
Claude Code starts in the terminal, but that is one door of several: the desktop app, the VS Code and JetBrains extensions, the cloud, Dispatch for background work, and the developer platform.
What AI Coding Agents Leave Out (and How to Put It in Your Prompt)
Five AI coding tools built 15 apps. Zero shipped security headers and only one had rate limiting. Here's what agents skip by default and how to prompt around it.
Your homepage is a login wall: why a great app nobody can find is still invisible
If your front door is a log-in box, search engines and AI assistants have nothing to index or cite. Here's the login-wall anti-pattern, why it's fatal for discovery, and the fix.
Where Should a Small Business Host Its Website and Apps? (Part 2 of 5)
A brochure site, a booking app, and a multi-client agency each want a different host. Here is how a small business picks among Cloudflare, Vercel, and Netlify without quietly getting locked in.
Taking Money as a Small Business: Square, Stripe, and Staying Portable (Part 3 of 5)
The processor you pick to take payments quietly owns your customer card data. Here is how a small business chooses among Square, Stripe, and PayPal and keeps the freedom to switch later.
RAG and Agents Over Your Own Documents, Without the Hype (Part 4 of 5)
Pointing AI at your own manuals and policies is cheap and genuinely useful, but only sometimes. Here is when a small business should build retrieval or an agent, and when a shared doc beats both.
Hire, Upskill, or Outsource Your AI Work? A Small Business Decision (Part 5 of 5)
Most small businesses do not need to hire a $200k AI engineer. Here is how to choose between hiring, training someone you already have, and outsourcing a defined project.
Local AI or Cloud APIs for a Small Business? The Honest Cost Math (Part 1 of 5)
For most small businesses the cloud AI bill is a rounding error, not a reason to buy a GPU. Here is the real local-versus-cloud math and the two cases where owning the hardware actually pays off.
The Free-First Home AI Hardware Ladder, From Your Old PC to a Real Training Rig
You do not need to buy anything to start training and running AI at home. Here is the honest hardware ladder for 2026, from free cloud GPUs to a used RTX 3090, with real prices, what each tier can run, and when to rent instead.
Going Frameworkless on Cloudflare, and Why You Still Reach for Vercel, Netlify, and Square
You can ship a real app on Cloudflare with no framework and almost no build step. Here is when that wins, where it bites, and how to split work across Vercel, Netlify, and Square without getting locked in.
Antithesis and the Quiet Comeback of Deterministic Simulation Testing
Antithesis runs your whole system inside a deterministic hypervisor so every bug reproduces perfectly. Here's how it works, who uses it, and the FoundationDB lineage behind it.
What $200k AI Jobs Actually Ask For, and How to Practice Every Skill Free
Cleared and commercial AI job posts converge on the same open-source stack. Here are the GitHub projects that move you up the pay band, what each does behind the scenes, and how to practice every one at home for free.
Obsidian: The Note App That Keeps Your Notes as Plain Files You Own
Obsidian stores notes as plain Markdown files on your own device, went free for commercial use in Feb 2025, and only charges for Sync ($4/mo) and Publish ($8/mo).
Streaming Pipelines and Platform Engineering: The Two Disciplines Behind Systems That Scale
Event streaming (Kafka, Kafka Streams, Flink) moves data reliably at scale; platform engineering (golden paths, shared libraries, IDPs) scales the teams that build on it. Here's how both work and where to learn them.
Your Home Insurance Outran Your Paycheck. Here's the Machine Behind It.
Home insurance rose faster than inflation and faster than most raises since 2019. The reason runs through tax-free Bermuda cat-bond vehicles and ends with you as the backstop. The honest data, sourced.
The Cloudflare Developer Platform, Explained Like You're Going to Ship on It
Cloudflare quietly turned its CDN into a full app platform: edge compute, SQLite, object storage, queues, AI inference. Here's the honest tour of what each piece does, what it costs, and where it bites.
Clerk: The Fastest Way I've Found to Ship Real SSO
Auth is the part of every app that quietly eats a week. Clerk turns sign-in, MFA, and even enterprise SAML SSO into a few components and a config screen. Here's the honest rundown.
Baseten, Explained: What a $13B Inference Platform Actually Does
Baseten raised $1.5B in June 2026 at up to $13B. Here's what its inference platform, Truss packaging, and Chains SDK actually do for people shipping models.
How to Build the Skills Behind a Lead AI/ML Platform Engineer
Senior AI/ML platform roles want agentic AI, distributed systems, Kubernetes, MLOps, and streaming, all at once. Here's the honest capability map, the sources I'd actually use, and projects that prove each one.
Gumroad Discover: Passive Sales From a 1.6M Product Marketplace
Gumroad Discover surfaces your digital products to 1.6 million+ listed products' worth of buyer traffic. Here's how to optimize your listings, use commission boosts, and leverage tags for passive discovery.
GLM 5.2: What Z.ai's Open-Weight Flagship Actually Ships (and Costs)
GLM-5.2 ships MIT-licensed open weights, a 1M-token context, and SWE-bench Pro 62.1. Here's what runs on your hardware, what it costs, and what's hype.
The machine-readable layer most B2B SaaS sites skip (and why AI assistants can't recommend you without it)
A new Mega Analyzer check for B2B SaaS and AI-product sites: SoftwareApplication schema, the Organization entity graph, FAQPage over your existing FAQ, per-page Open Graph, and breadcrumbs.
Hotels ship the schema and forget the listings: the booking surface most hotel sites leave unbound
A new Mega Analyzer check for hotels: Hotel/LodgingBusiness schema, HotelRoom + Offer, and the local-listing sameAs graph (Google Hotels, Apple Maps, Yelp, TripAdvisor, OTAs) that most hotel sites leave unbound.
Quantum Did Not Just Kill the Trillion Dollar Data Center Bet. Here Is What Actually Happened.
A widely shared Medium piece says one quantum result invalidated AI's energy buildout. The peer-reviewed paper is real. The supremacy framing is contested. Here is what AI builders should actually take from it.
Running Your Own AI On-Prem in 2026: What It Actually Costs, Where It Wins, Where Cloud Still Beats It
An honest cost, speed, privacy, and IT-ops contrast between cloud LLM APIs and on-prem inference. With the real Apple Silicon and RTX 5090 numbers, the per-token price tables, the break-even math, and the security warnings most local-AI posts skip.
Stop Prompting, Start Designing Loops. The Honest Version Of A Tweet That Set Off Coding Twitter.
Peter Steinberger said you should be designing loops, not prompting agents. Boris Cherny and Addy Osmani backed the idea. Here's the practitioner read, the unsexy cost problem nobody is naming, and how it maps to /loop, /schedule, skills, subagents, and hooks.
Publishing Book Tools on GitHub: DA 95 Backlinks for Free
GitHub has a Domain Authority of 95 and is one of the most-cited sources in ChatGPT responses. Here's how publishing open-source tools related to your book generates permanent backlinks and AI visibility.
Karpathy's Four AI Coding Rules Are the Floor, Not the System
I tested the popular Karpathy-guidelines CLAUDE.md against a stack with 300+ shipped tools. Great on day one, net negative by month two. Here's the earned version.
Crunchbase + Wikidata: The Two Free Profiles That Feed Google's Knowledge Graph
Crunchbase and Wikidata are primary data sources for Google's Knowledge Graph. Here's how creating free profiles on both platforms can trigger a Knowledge Panel for your brand or personal entity.
Claude Fable 5 Suspended Days After Launch: a US Export-Control Order Pulled It for Everyone
A US export-control order pulled Claude Fable 5 days after launch. Here's what to use instead.
What Shipped Since You Installed Claude Code: Opus 4.8, Dynamic Workflows, The Agent View, And The Q2 2026 Surface
A field guide to what landed in Claude Code in spring 2026: the Opus 4.8 default and fast mode, dynamic workflows, the claude agents view, background sessions, /goal, the /simplify to /code-review split, remote control, and how Codex kept pace.
One Diocese, Many Parishes, One Template
A diocese has dozens or hundreds of parish sites, each an island, each a different cost. Here is how to run them all from one shared template, host them free, and consolidate the bills into one subscription per layer.
Your Self-Storage Site Lists One Set of Hours. Renters Need Two.
The FTC can fine fake-review violations up to $51,744 each, yet most self-storage sites still hide gate-access hours and bury move-in special terms. Here is the fix list.
Why Most Laundromat Websites Lose the 'Open Now' Search (and the Checks That Win It Back)
Laundromats live and die by 'laundromat near me' and 'open now.' Here are the schema, hours, payment, ADA, and FTC checks that decide whether search and AI can find you.
Your Car Wash Membership Auto-Renews. Does Your Website's Cancel Flow Follow the Law?
Unlimited wash clubs auto-renew, and the cancel flow is now a legal surface. Here are the schema, membership, review, and ADA checks your car wash site needs to pass.
What a craft butcher or DTC meat brand has to make machine-readable before AI will sell it
The USDA, FTC, and FDA rules plus the structured-data, perishable-shipping, and provenance moves that make a small meat brand legible and trustworthy to AI shopping.
Why Paying For A Website Platform Does Not Mean Your Site Is Actually Built Right
A monthly website subscription buys you a baseline, not modern structured data, security, accessibility, or AI-answer readiness. Here is a 20-minute self-audit...
The SpotHopper Website Checklist for Restaurant Owners: What to Check and What to Fix
If SpotHopper builds and hosts your restaurant site, you can fix the meta description and page title yourself, but missing structured data, multiple H1 tags, and...
The Popmenu Checklist For Restaurant Owners: What To Verify Before And After You Sign
Popmenu ships Restaurant schema and runs your site for you. Here is what an owner still has to check by hand: menu crawlability, pop-up overuse, lock-in...
On HubSpot CMS? Here Is What to Check Yourself Before You Pay Anyone
HubSpot Content Hub builds a fast, tidy site, but schema markup is not automatic and the consent banner is off by default. Here is the owner-operator checklist...
These Are Not Church Problems. They Are Small Business Problems.
I audited a stack of parish websites and found the same seven gaps every plumber, bakery, and HVAC company has. Missing schema, a mis-pinned map listing, no email auth, blocked AI crawlers. All fixable in a weekend, mostly free.
HVAC companies hide their best trust signals: the Mega Analyzer's HVAC-readiness checks
Most heating-and-cooling companies have the credentials and reviews to win local search, but their websites hide all of it. Here is what the Mega Analyzer's HVAC-readiness layer looks for.
Why Your Church Is Invisible to ChatGPT, and the Free Fix
A parishioner asks an AI assistant for Mass times and gets a stale answer or nothing. Here is why it happens, drawn from real audits, and the free fixes any parish staff can do this week.
Google Discover: The Passive Traffic Firehose Most Sites Ignore
Google Discover sends content to 800 million+ devices based on interest graphs, not keywords. Here's how to optimize for Discover's unique ranking signals and start generating passive traffic.
A Job That Took Two and a Half Hours Now Takes Thirty Minutes. One Change.
A scraper was running up to 150 minutes a pass, seven times a day, and the bill kept climbing. The fix was a dozen lines: stop waiting on one thing at a time.
The Mega Analyzer now reads men's-health and TRT clinics: the YMYL trust signals a hormone clinic has to ship before AI will cite it
A new Mega Analyzer vertical for men's-health and TRT clinics. MedicalClinic schema, the empty-address and shared-@id multi-location traps, credentialed authors, FAQPage, reviews, AI-crawler access, and the FDA/Ryan Haight/FTC/HIPAA lines.
Electrician websites are leaving their best asset invisible: the Mega Analyzer's electrician-readiness checks
Most licensed electricians have the reputation and credentials to win local search, but their websites hide all of it. Here is what the Mega Analyzer's electrician-readiness layer looks for.
The Mega Analyzer now reads handyman and home-services sites: what a local trade has to ship before 'near me' can find it
A new Mega Analyzer vertical for handyman and home-services businesses. LocalBusiness schema, Google Business Profile, the client-rendered-SPA trap, placeholder phones, reviews, and the trade-licensing claims that have to be substantiated.
The Mega Analyzer now reads DTC stores: what an online shop has to make machine-readable before AI will sell it
A new Mega Analyzer vertical for online stores. Product/Offer completeness, country-of-origin, reviews-as-data, FAQ, and the entity graph that decides whether AI shopping can find, trust, and sell your products.
Cross-platform brand presence: the sameAs breadth that turns a young brand into one AI can cite
AI answer engines reconcile a brand from the breadth and consistency of its profiles. A practical, no-budget playbook for the cross-platform presence that makes a young store look established to machines and humans.
I Cut a Recurring AI Bill by More Than Half in an Afternoon, Without Touching Quality
The model bill on a project I run was about 140 dollars a month. An afternoon of looking, not rewriting, took it under 25. Five checks that find waste in any recurring AI spend.
Your AI Is Quietly Making Up the Numbers in Your Reports
An AI report on my own news site claimed to count 163 stories but had actually counted 52. The fix is simple: let the model judge, and compute every number in code yourself.
An RIA's Advisors Are the Product — and 0% of Their Bios Have Schema
Wealth managers run generic Organization schema with zero sameAs and no Person markup on the advisors who are the firm. Here's the fix — and why review schema that helps a home inspector is a compliance risk for an RIA.
The New-Home Builder's Website Is Invisible to AI Shoppers — Here's the Fix (and the Compliance Traps)
Production builders ship great photos and thin schema. Price and move-in-ready availability are invisible to AI. Here's the structured-data fix, plus the Fair Housing, RESPA, and Reg Z traps.
Your Home-Inspection Site Has 6,000 Reviews the AI Can't Read — Here's the Fix
Home inspectors ship the richest local-business schema in their category and still lose the AI citation. The missing layer is FAQPage, structured reviews, and three validity defects. Here's the deterministic fix.
BigLaw Readiness in the Mega Analyzer: The At-Scale Checks a Single-Page Legal Audit Misses
A new Mega Analyzer layer fires on large multi-office law firms and checks the things single-page legal audits can't: corpus-wide Article schema, Chambers/Legal 500 sameAs, Event and multimedia markup, and hreflang against the real office list.
Aircraft Lessors Are Entity-Dark: Your Dealmakers Have No Schema, and That's the Whole Problem
Aircraft asset managers ship a bare WordPress org node and no Person schema on the people who ARE the firm. Here's the machine-readability fix for B2B aviation finance, plus the fund/ABS disclaimer trap.
The Domain You Don't Send Email From Is Still Spoofable: Locking Down Parked Domains
A domain with no email is the easiest one to forge, because nobody set a policy saying 'reject everything.' Two DNS records close the hole. Here's the exact lockdown for a parked domain.
The Outdated JavaScript Problem: Why Old Libraries Are the Most Common Unpatched Hole on the Web
A jQuery from 2019 or a Bootstrap from 2018 sitting on your site is a known, public, exploitable bug with a CVE number attached. Here's how to find every outdated library you're running and upgrade it.
Generative Engine Optimization: Why Traditional SEO Is No Longer Enough
Traditional SEO optimizes for ten blue links. Generative Engine Optimization targets the AI models that now answer questions directly. Here's how citation farming works differently for ChatGPT, Perplexity, and Google AI Overviews.
How to Send Your Own Email Digest From Serverless Functions (Resend and the Alternatives)
Skip the newsletter platform. A subscriber store, one scheduled function, and a sending API give you double opt-in, one-click unsubscribe, and a daily digest for almost nothing a month.
The Mobile-Presentation Parameters Modern Sites Still Get Wrong
Five mobile-presentation parameters that ship broken on otherwise-modern responsive sites. Each fix is one to three lines of CSS. Validated against an iPhone, Pixel, and iPad probe.
12 non-profit transparency checks the Mega Analyzer now runs (when it detects a 501(c)(3))
When the Mega Analyzer detects a 501(c)(3), it now runs twelve donor-trust checks gated on the vertical: EIN visibility, GuideStar links, donation platform freshness, NGO schema, DonateAction, more.
Legal-Site Readiness in the Mega Analyzer: What an Am Law Audit Catches Before the Ethics Committee Does
A new Mega Analyzer section runs only on law-firm pages and flags missing inline advertising notices, prior-results disclaimers, Person/Attorney schema, and EU cookie banners against bar-rule and schema.org references.
Defense-Contractor Readiness in the Mega Analyzer: Why Federal Procurement Officers See a Different Site Than Marketing Thinks They Do
A new Mega Analyzer section that runs only on US defense-contractor pages and flags missing CMMC level, DFARS 252.204-7012, NIST SP 800-171, ITAR / EAR, FedRAMP, SAM.gov / CAGE / UEI / NAICS disclosure, Section 889, and Capability Statement PDF against the DoD, FAR, DFARS, NIST, and State Department references.
CPA-Firm Readiness in the Mega Analyzer: Why a Generic SEO Scan Doesn't See AICPA Peer Review, Circular 230, or FTC Safeguards Rule Gaps
A new Mega Analyzer section that runs only on CPA / accounting-firm pages and flags missing AICPA peer review references, Circular 230 acknowledgment, FTC Safeguards Rule statements, and Person+Accountant + AccountingService schema against AICPA, IRS, and Idaho/state-board references.
12 brand and copy hygiene checks that should be in your SEO audit (now in the Mega Analyzer)
Twelve audit patterns that silently cost visibility and trust, including a typo that looks correct to the eye but reads wrong to every search engine. All shipped to the Mega Analyzer.
Robots.txt as a Platform Fingerprint — Catching Pixieset, Squarespace, and Showit When CDN Detection Misses
URL-regex platform detection misses any SaaS that lets tenants serve assets from a custom domain. The Mega Analyzer now reads robots.txt as a fallback fingerprint — catches Pixieset, Squarespace, and Showit when CDN URLs are masked.
Photographer Sites Need ImageObject and Photograph Schema — Auto-Detected by Vertical
Photographer marketing sites need ImageObject and Photograph JSON-LD with copyrightHolder, license, and acquireLicensePage fields, but the existing audit only fired when those blocks already existed. Round 8 detects the gap by vertical.
Vercel BotID and Cloudflare Bot Fight Mode Silently Block AI Search — Now Auto-Detected
GPTBot, ClaudeBot, PerplexityBot, and Google-Extended all hit the same wall when a marketing site enables Vercel Attack Challenge or Cloudflare Bot Fight Mode. The Mega Analyzer now flags it as a critical finding with platform-specific remediation.
Mega Analyzer goes Shopify-aware — six new detectors and one suppressor
Auditing a Shopify storefront in a generic SEO checker drowns the merchant in false positives: visually-hidden H1s, http schema.org context, paired og:image http+secure_url, /policies/* paths the checker misses. The Mega Analyzer now fingerprints Shopify and applies a Shopify-aware rule pack — six new Shopify-only detectors and one suppressor that fixes a long-standing /policies/* false negative. Patterns generalized from a real audit; brand details anonymized.
Content Credentials: How to Prove You're Human in an AI-Generated World
The C2PA standard lets you cryptographically prove that your content was created by a human. With the EU AI Act requiring AI content labeling, Content Credentials are becoming a trust signal. Here's how to implement them.
Ontology Just Beat Vector Embeddings. Here Is What That Means for Your Small Business.
Vector search was the default way AI tools read your business data through 2024. Graph shaped knowledge structures are the default in 2026. Here is what changed, why it matters at 5 person or 50 person scale, and a two CLI workflow to migrate your own data using Claude Code and Codex.
ProfilePage.mainEntity as a bare @id reference — the Search Console error that hides in valid schema
Search Console flags ProfilePage and AboutPage with 'Invalid object type for field mainEntity' when the referenced Person/Organization lives on another URL. The schema is valid. The Rich Results Test is green. Here is the fix and why it is necessary.
Date-only YYYY-MM-DD in JSON-LD trips Search Console — the ISO 8601 fix
Schema.org accepts Date or DateTime on dateCreated, dateModified, datePublished. Google Search Console's strict validator emits 'Invalid datetime value' for the date-only form. Convert to ISO 8601 datetime with timezone offset. Here is how and why.
Spot the Spread: How to Find the Money Hiding Inside Your Own Business
Quant funds make billions exploiting structure, not direction. A small business owner can apply the same mindset, look for the spreads inside their own operation, and find money they did not know they were leaving on the table.
If You're Going to Have AI Write Your Business Tools, Use Python
A public benchmark gave Claude Code the same task in thirteen programming languages and measured the result. Python, Ruby, and JavaScript were 1.4 to 2.6 times faster and cheaper than statically typed alternatives. Here is what that means for any small business hiring AI-assisted code work.
The Case for an AI That Remembers You: Why Persistent Memory Beats a Fresh Prompt Every Time
An open-source agent called Hermes crossed 110,000 GitHub stars in ten weeks because of one feature: it remembers what you taught it last time. Here is what that changes for a small operator who runs the same workflow week after week.
The $50-a-Month AI Stack: How a Small Business Can Automate the Boring Half of the Work
A SitePoint engineer runs his full multi-platform social posting pipeline for under $50 a month using Claude and n8n. The same architecture maps onto invoicing, monthly reports, and customer follow-ups for any small business.
Your Small Business Probably Runs on WordPress. The 15-Minute Security Audit Most Owners Have Never Done.
43% of the internet runs WordPress. Most small-business WordPress sites haven't been audited since the day they were built. Here's the 15-minute self-audit, the seven checks that catch 90% of real risks, and the free tool that runs them for you.
Why the Claude Skill Linter Exists (and the Fourteen Patterns It Catches Before You Ship)
An engineering team lost $47,000 because two AI agents got stuck in a loop for 11 days. The patterns that caused it are visible in the skill text. The Skill Linter catches 14 of them in a single paste.
Why the AI Vendor Cost Reverse-Calculator Exists (and the 60-Second Negotiation It Powers)
73% of funded AI startups are wrappers reselling OpenAI or Claude. The Reverse-Calculator gives you the actual API cost behind the vendor's quote in 60 seconds, plus the negotiation prompt that closes the gap.
Before You Pay an Agency $3,500 a Month for 'Proprietary AI,' Open Your DevTools
An engineer reverse-engineered 200 funded AI startups. 73% are reselling OpenAI or Claude with a custom UI. Here's the 30-second test that catches the markup, the four red flags, and the honest math on what you should actually pay.
Why Google Rich Results Test only shows FAQ and Breadcrumb on your About page
Your schema validates clean but Google Rich Results Test returns only FAQ and Breadcrumb — or nothing at all. It's not a bug in your schema. It's a schema-type-eligibility issue Google does not document well, and the fix is one new schema type.
Before You Hire an AI Agency: A 30-Minute Self-Audit (Part 1 of 2)
Most AI agency pitches collapse under a 30-minute self-audit you can run with no technical skills. Eleven specific questions, the F12 check, and the cost math that lets a 25-person company evaluate any AI vendor without hiring a consultant to evaluate the consultant.
After You Fire the AI Agency: The 30-Day Migration Playbook (Part 2 of 2)
You ran the 11-question audit and decided to leave the agency. Here's the 30-day plan to extract your data, rebuild the workflows on Claude Pro for $20/month, and keep the lights on during the transition. Specific checklist, real cost math, no fluff.
How a 12-Person Trades Business Can Sell Its Permit-Pulling Cheat Sheet as a $19 PDF This Weekend
You already know things customers would pay for. Most small-business owners never package them. Here's the six-step weekend version: pick the workflow, draft it in Canva, ship it on Gumroad for free, link it from the page you already have. No funnel. No list.
The Invoice-Chase Skill, Opened Up: What's Inside Anthropic's Launch-Hero Workflow and How to Bend It to Your Business
Anthropic's invoice-chase skill became the marketing hero of the Claude for Small Business launch. Here's the instruction template behind it, the eight failure modes most owners hit in the first week, and the customizations that matter for service vs. product businesses.
Why five inline Person blocks broke my AI Search visibility (and the fix)
JSON-LD merges entities by @id. Declare the same Person five times inline with conflicting alternateName values and Google AI Mode picks Amazon Author Central instead of your own site. Here's how I caught it and what to do.
GDPR says people must be able to change their mind about cookies. Most consent banners do not let them.
A working consent banner is half of the legal job. The other half is a footer link that reopens it. Here is what the law actually requires and three ways to wire it up.
The Connector Permission Cheat Sheet for Claude for Small Business: Twelve Connectors, Twelve Minimum-Permission Sets
When you connect QuickBooks, Gmail, Stripe, or Slack to Claude, the default OAuth scope is usually broader than what the skill actually needs. Here's the smallest viable permission set for each of the 12 launch connectors, with the rationale.
Claude for Small Business Just Shipped. Here's a Plain-English Walkthrough for the 25-Person Landscaping Company.
Anthropic released Claude for Small Business on May 13, 2026 with 15 ready-made skills for payroll, books, invoicing, marketing, and onboarding. Here's what's actually in the box, how to turn it on this week, and the under-$100 stack it fits into.
Wall Street Is Betting on the Wrong AI Layer. So Are Most Small Businesses. Here's What That Means for Your Margin.
Wall Street pours money into AI models. SMBs pour money into AI vendors. Both are buying the layer that's commoditizing fastest. The boring layer underneath (skills, connectors, observability, cost discipline) is what actually decides whether AI makes you money.
Nine AI Mode entity-binding bugs that pass every schema validator
Nine schema patterns that validate clean, index without issue, and still cost a site its Google AI Mode citation. What to look for and how to fix each one.
How a Small Business Runs AI Agents Without a $47,000 Surprise Bill
An engineering team woke up to a $47K AWS bill because two AI agents got stuck in a loop for 11 days. Here's the cheat-sheet version for small-business owners: the four safeguards that prevent 99% of agent disasters, and the honest answer to whether you should run agents at all yet.
Your Coverage Report And Your Sitemap Disagree. Here's How To See Where.
Export three CSVs from GSC, Bing Webmaster, and your sitemap. Paste. See every URL that's in the sitemap but not indexed, every URL that's indexed but not in the sitemap, and every ranking drop month-over-month. No OAuth, no paid rank tracker.
Gravatar is the indie operator's Wikidata, and most people miss it
Wikidata feeds Google's Knowledge Graph. So do Gravatar, ISNI, VIAF, Linktree, and a handful of industry-specific directories most operators never claim. The full portable identity-anchor playbook for people who can't qualify for Wikidata.
Your schema validates. Google AI Mode still can't find you. Here's the missing piece.
A site I work on dropped out of Google AI Mode citations for the operator's own name after a routine schema cleanup. The fix was not more schema. It was binding the entity to Google's Knowledge Graph with kgmid, Maps cid, and dual-typed JSON-LD.
Your robots.txt Says One Thing And Your CDN Says Another
You disallowed GPTBot in robots.txt. Your CDN injects X-Robots-Tag: all. Your ai.txt was never created so it defaults to permissive. Three sources of truth, three different answers, and the crawler picks whichever it likes. This tool flags every disagreement per bot.
Your AI Bot Policy Lives In Three Files, And They Probably Disagree
Your robots.txt disallows GPTBot. Your ai.txt allows OpenAI-User. Your meta robots says index,follow. Crawlers pick the most permissive interpretation and you accidentally opt into training you thought you'd blocked. One tool emits all three files, consistent, per 22+ named AI crawlers.
Why Your Analytics Are Lying: How Ad Blockers Hide 40% of Your Visitors
Ad blockers, privacy browsers, and tracking prevention tools hide up to 40% of your real traffic from Google Analytics. Cloudflare Zaraz fixes this for free. Here's how to set it up.
Self-host the CDN-hosted fonts and scripts. The DNS lookup costs more than the asset.
Public CDNs add 100-300ms of cold-start latency, leak visitor IPs, and break when the CDN goes down. The fix is the same five-step recipe whether the asset is a font, a CSS framework, or a small JavaScript library.
Your Article Covers The Topic But Doesn't Answer The Question
Headings Outline validates structure. Heading Gap Audit compares to competitors. This tool compares your headings to your actual target query — catching the 'good article, wrong question' failure mode. Per-heading alignment score with a semantic-expansion pool for related-query variations.
Your Lighthouse Score Is Green But Real Users Are Suffering
Lighthouse tests a single cold load with throttled network on a simulated device. CrUX aggregates 28 days of real-user sessions. When they disagree — lab green, field red — your synthetic test has a blind spot. The tool queries Google's PSI API for both, computes the delta per metric, and emits the diagnosis prompt.
Cut The Filler Paragraphs — Measuring Information Payload Per Paragraph
A 2,000-word article with one dense paragraph and twenty filler paragraphs ranks worse than a 600-word article that's dense the whole way through. This scorer measures information payload per paragraph — noun-phrase density, entities per 100 words, specific-claim markers (numbers, dates, money), filler ratio — and flags paragraphs to cut.
Circular Canonicals Tank Both Pages — Canonical Cluster Mapping
Per-URL canonical checks catch the obvious bugs. Graph-level canonical mapping catches the subtle ones — circular canonicals (A→B, B→A), canonical pointing at a noindex page, cross-protocol drift. The mapper fetches your sitemap, builds the target graph, clusters by canonical destination, and flags every conflict type.
llms-ctx.txt and llms-ctx-full.txt: The FastHTML Extensions to llms.txt
The pair of files that take llms.txt one step further by pre-wrapping linked pages in XML tags ready for one-shot LLM prompt injection. Worth it for docs and KB sites, skippable for the rest.
The HCU Classifier: Why One Bad Page Can Drag Down Your Whole Site
Google's Helpful Content System runs continuously inside core ranking and applies a sitewide demotion when unhelpful signals stack up. Here's how to detect it and what to fix.
The Cloudflare Agent Readiness Score and What It Actually Checks
isitagentready.com tests for 13+ signals across discoverability, content, bot access, and capabilities. Here's what each one does, what's worth shipping, and what's optional for most sites.
AGENTS.md: The Root-Level README That AI Coding Agents Actually Read
Cursor, Claude Code, Codex, Copilot, Gemini CLI, and 30+ other coding agents look for /AGENTS.md before they touch your repo. Here's the spec, why it matters, and what to put in it.
Thirteen Probes For The Agent-Ready Web, And A Browser-Side Tool That Runs All Of Them On Any URL
A new free auditor at jwatte.com runs the same thirteen probes Cloudflare and isitagentready.com use to score a site as agent-native: Link headers, Markdown negotiation, Content Signals, Web Bot Auth, API Catalog (RFC 9727), MCP Server Card (SEP-2127), Agent Skills v0.2.0, WebMCP, OAuth metadata, x402 plus the three commerce protocols. Browser-side, no signup, no data stored.
WebMCP — The First Real Agent API For Your Site (And Why llms.txt Wasn't It)
Chrome shipped WebMCP in Canary on February 10, 2026. It's the first browser-native protocol for sites to expose agent-callable actions, and it's what llms.txt was supposed to be — but actually shipping. Two surfaces (Declarative API + Imperative API), the four declarations to add, and why getting in early on a W3C Community Draft is the right move for SMB sites this year.
Every Unsplash Photo On Your Site Legally Needs Attribution
Unsplash's license requires attribution. Pexels' license requires attribution. Free stock doesn't mean no strings. Beyond legal exposure, missing image credit disqualifies you from Google's image-rich-results (requires ImageObject.creditText or copyrightNotice). The audit checks every image for attribution + C2PA manifest + ImageObject schema.
FAQPage Schema Alone Won't Get You Featured — Answer Quality Does
Your FAQPage JSON-LD is valid. Google still isn't using your answers as featured snippets or AI citations. The reason: answer quality. An answer opening with 'It means...' without naming what 'it' is doesn't stand alone. The scorer checks 8 linguistic signals known to predict which answers get lifted.
Three Agent Manifests Your Site Could Ship In 2026 — And Which One Actually Matters
Three different well-known files claim to be the agent's entry point to your site: llms.txt, /.well-known/ai-plugin.json, and /.well-known/webmcp. They do different jobs, have very different adoption, and only one is worth shipping first in May 2026. The honest comparison.
"Host Did Not Return Markdown Content When Accept: text/markdown Was Requested" — What That Means And How To Fix It
Run the Agent Runtime Readiness audit and you may see this warning on the third test. It is the audit telling you AI runtimes are sending Accept: text/markdown to your URL and getting full HTML back instead of a slim markdown twin. Here is what the warning is actually testing, and the two fix paths — the Cloudflare toggle and the origin-side content negotiation pattern.
Fix The Markdown For Agents Warning On Vercel — Edge Middleware Pattern
If your site is on Vercel and the Agent Runtime Readiness audit warned about missing Markdown content negotiation, here is the middleware.ts pattern that fixes it. Two patterns — companion file and runtime conversion — with route matchers and the Vary header your CDN needs.
Squarespace, GoDaddy, And The Markdown For Agents Warning — What You Can And Cannot Fix
Closed website builders like Squarespace and GoDaddy Website Builder cannot fully fix the Markdown for Agents warning because they don't expose HTTP-level controls. GoDaddy Managed WordPress and cPanel hosting can. Here is the honest split — what's possible on each, the partial workaround for closed platforms, and the real .htaccess + PHP fix for managed WordPress.
Fix The Markdown For Agents Warning Without A CDN — Nginx, Apache, And Caddy Configs
If your site is served directly from Nginx, Apache, or Caddy with no CDN edge layer in front, you can still implement the Markdown for Agents content-negotiation pattern. Three side-by-side configs — what to add, where to add it, and how to test it works.
Fix The Markdown For Agents Warning On Netlify — Edge Functions Pattern
If you saw "Host did not return Markdown content" from the Agent Runtime Readiness audit and your site is on Netlify, the fix is a 30-line Edge Function that intercepts Accept: text/markdown and serves a clean markdown twin. Two patterns — companion file and runtime conversion — with the netlify.toml route config.
Fix The Markdown For Agents Warning On Fastly — Compute@Edge Pattern
If your site is on Fastly and the Agent Runtime Readiness audit warned about missing Markdown content negotiation, here is the Compute@Edge JS pattern that fixes it. Why VCL alone cannot do this, the full handler, and the cache configuration that lets the edge keep both response shapes.
Fix The Markdown For Agents Warning On AWS CloudFront — Lambda@Edge Pattern
If your site is on AWS CloudFront and the Agent Runtime Readiness audit warned about missing Markdown content negotiation, the fix is a Lambda@Edge function on the viewer-request or origin-response trigger. Here is the implementation, the deployment path, and why CloudFront Functions cannot do this on its own.
When the auditor can't reach your site: save your page or briefly unblock the probe
Step-by-step browser walkthrough (Chrome, Edge, Firefox, Safari) to save your rendered HTML, plus exact menu paths to allow the probe through the top 10 hosting and bot-protection platforms. Beginner-level.
E-E-A-T Is Content. Trust Signals Are Infrastructure.
Google News Initiative's Trust Project defined 12 structural artifacts that signal 'a real organization stands behind this content' — editorial policy, corrections policy, author bio, physical address, external review platforms. The audit checks all 12 in one pass. Distinct from E-E-A-T content audits; this is infrastructure.
Why Your Blog Has 200 Posts But Only Page 1 Is Indexed
Google deprecated rel=prev/next as a ranking signal in 2019, but it still uses the chain for discovery. More urgently: the canonical-to-page-1 anti-pattern — where every paginated page declares /page-1/ as canonical — silently de-indexes every deep page. The checker follows the chain up to 10 pages and validates every rule.
I Built An MCP Server Recommender Because Configuring Claude Desktop Should Take Five Minutes, Not Five Hours
A new free generator that takes a checklist of what your team uses and emits a paste-ready claude_desktop_config.json with the right Model Context Protocol servers wired up. Vendor-published servers labeled green, community labeled amber, reference labeled blue. Covers GitHub, Linear, Atlassian, Slack, Notion, Stripe, Postgres, Sentry, and more.
I Built A Hedge Language Auditor After Reading Kevin Indig's Citation Research. Here Is What It Catches.
A new free browser-based tool that scores any article on the four research-backed AI citation frameworks: BLUF, definite phrasing density, entity density, strategic repetition. Surfaces every hedge word, every BLUF-fail section, and the repetition gaps that hide your strongest claim from retrieval.
I Built A Printable Fraud Reflex Card After Reading The FBI's 2025 Crime Report. Print One Per Workstation.
A new free generator that emits a printable one-pager of the six AI-era fraud reflexes — callback rule, dual-channel verify, did-I-initiate-this, voice-print phrase, deferred-response wait, clean-channel review — customized per business with internal contact details. Grounded in the FBI's 2025 IC3 report ($20.9B losses, $893M AI-enabled).
I Got Tired Of Hand-Writing Cloudflare Cache Rules For AI Bots. So I Built A Generator.
A new free generator that emits a paste-ready Cloudflare Cache Rule expression for AI bot HTML traffic, with TTL bands by content type and dynamic-surface exclusions. Three output formats: dashboard expression, Terraform, cURL/API. Reduces 499 timeouts that hide your pages from ChatGPT, Claude, and Perplexity.
Three Tests Every Page Has To Pass For An AI Agent Runtime, And A Tool That Runs Them
A new free auditor that runs the three tests every page should pass for AI agent runtimes (Cloudflare Project Think, OpenAI Agents SDK): server-rendered content ratio, inline JSON-LD presence, Markdown-for-Agents content negotiation. Backed by the April 2026 runtime-layer shift.
I Built An Ad Library Deeplinker Because Investigating One Brand Across Five Platforms Was Eating Mornings
A new free utility that takes one brand or Page name and produces paste-ready deeplinks across Meta Ad Library, Google Ads Transparency Center, X Transparency, LinkedIn Ad Library, TikTok Commercial Content Library, and EU DSA repositories. No scraping, no login, just the canonical search URLs.
The Retail Quant Stack: Coinbase, ccxt, vectorbt, And A Working Momentum Backtest
A specific, free or near-free GitHub stack a retail trader can run on a laptop: ccxt for data, vectorbt for backtesting, freqtrade for live execution, coinbase-advanced-py for direct API access. Includes a runnable time-series momentum backtest, realistic Coinbase fees, and the academic results to expect.
Why Top Quant Desks Outperform, And How U.S. Tax Code Multiplies The Edge
Renaissance, Citadel Securities, Jane Street, XTX, HRT, DRW: a tour of who actually makes the money, why their edge persists, and how Section 1256, Section 475(f), and the post-Lender family office structure quietly compound that edge inside U.S. tax code.
Inside Quadrature Capital and Radix Trading: What's Actually Public, And What Retail Can Borrow
Two of the most secretive AI quant firms in the world, Quadrature in London and Radix in Chicago, leak just enough about their stack to be instructive. Here is what is actually verifiable about how they trade, and the parts a retail investor with a laptop and a Coinbase account can reasonably copy.
MyClaw, Microsoft, And The Honest Answer For Small Operators Who Want An AI Assistant Without Hiring An Employee
MyClaw is managed hosting for the OpenClaw agent, priced from $16 a month. The product is real, the productivity case is real, and the security disclosures are thinner than I would want. Here is what it actually does, where it fits, and when an SMB operator should pick it over Claude.ai direct or Microsoft 365 Copilot.
$5.4 Billion And 17.5 Million Ads: What The Meta Ad Library Actually Tells You About How U.S. Political Money Flows
I downloaded the Meta Ad Library's lifelong U.S. dataset for ads about social issues, elections, and politics. Here is what the real numbers say about who spends, where the money lands, and how political advocacy hides behind brand names. Plus comparable transparency reports from Google, X, LinkedIn, and TikTok.
The Best MCP Servers By Industry: A Practical Tour Of github.com/mcp For 2026
GitHub's MCP Registry now lists thousands of Model Context Protocol servers. This is a curated, cited tour of the ones that actually move work in finance, dev, support, marketing, and infrastructure, with copy-paste configuration examples for Claude Code and Claude Desktop.
Google Preferred Sources Is Not A Trust Algorithm. It Is A Subscribe Button. Here Is What To Do With It.
Google quietly expanded Preferred Sources globally on April 30, 2026. The feature lets users mark publishers they want to see more of in Top Stories. John Mueller has confirmed it is not an override of quality signals. Here is the publisher's playbook for actually getting selected, and where it fits with EEAT and the new AI search layer.
How to Verify an Auditor's Findings Without Trusting the Tool: Ethical Scraping for Small Operators
When an automated WCAG, cookie, or security audit flags your site, you need a way to reproduce the finding before you spend money fixing the wrong thing. A practical guide to scraping in a way that respects ToS, the CFAA, and the people running those sites.
The FBI Logged $20.9 Billion In Internet Crime Losses Last Year. Here Are Six Reflexes Small Business Owners Should Have Already Built.
The FBI's 2025 Internet Crime Report logged $20.9 billion in losses, with $893 million directly attributed to AI-enabled fraud and government impersonation losses nearly doubling. Here is the working set of reflexes a small business owner can build this week, grounded in the FBI numbers, neuroscience, and the new agent infrastructure that scaled the attack surface.
Why ChatGPT Cites Your Competitor And Not You: Four Writing Frameworks With The Research To Back Them
Four writing patterns that consistently improve the odds an AI model picks your page as the citation for an answer. Each is grounded in published research on how transformer attention works in retrieval, and each is testable on a single page in under five minutes.
The Conversation Has Moved Past The Model. The Conversation Is Now The Runtime.
On April 15, 2026, Cloudflare's Project Think and OpenAI's updated Agents SDK both shipped on the same day, both answering the same question: how does a long-running AI agent actually run in production. The runtime is the new browser, and your website is going to be evaluated against it, not against any individual model. Three tests to run before next week.
Your Site Might Be Invisible To ChatGPT For A Reason That Never Shows Up In Google Search Console
AI search runs in real time. If your origin doesn't respond fast enough, the agent gives up and your page is excluded before passage selection even starts. The signal lives in your edge logs as a 499 status code, and most SEO tools never look at it. Here is what to check, and what to fix.
/ultrareview: A Multi-Agent Code Review That Runs in the Cloud While You Keep Working
Claude Code v2.1.86 added /ultrareview, a fleet of reviewer agents that run in a remote sandbox and verify every finding before reporting it. Here is how to use it, what it costs, and how it compares to Cursor Bugbot, GitHub Copilot review, CodeRabbit, and Gemini Code Assist.
The Sitemap That Claims Every Page Was Updated Today — Fake-Fresh Detection
Many CMS plugins stamp every URL's lastmod as today on every sitemap rebuild. Google correlates the claim against the Last-Modified server header and the visible dateModified in JSON-LD. When sitemaps lie, Google eventually distrusts the whole freshness signal — and every page pays the cost.
A Deploy Silently Blocked GPTBot Last Week And Nobody Noticed
robots.txt gets regenerated on every deploy. ai.txt drifts when an agency 'cleans up' entries. llms.txt gets reformatted by the new CMS plugin. Any one of these can silently block an AI crawler and cost you citations for months before anyone notices. The diff tool compares two snapshots and classifies every change.
Paperclip: When Your AI Agents Need an Org Chart, Not a Prompt
Paperclip is an open-source platform that treats your AI agents like employees: org charts, monthly budgets, heartbeats, and tickets. Here is how it works across Claude Code, Codex, Cursor, and any agent that can receive a webhook.
Why a New Website Doesn't Show Up in Search, and the Free Stack That Fixes It
Sitemap submission, IndexNow, Google Business Profile, citation consistency. The discoverability stack I run on every new site, in priority order, with sources and the ranges of time each step actually takes.
How to Vet an HVAC Company: An 11-Point Scorecard That Saves Thousands
HVAC overcharging is structural, not occasional. $13K water heater quotes and $25K furnace pushes are documented in BBB complaints. Here's an 11-point scorecard you can use anywhere in the U.S.
BIMI and MTA-STS: Ship Both in an Afternoon, on Any Host (No $1,500 VMC Required)
BIMI puts your logo in the inbox. MTA-STS stops attackers from TLS-downgrading your incoming mail. Both use a DNS record plus an HTTPS-hosted file. Setup takes about an hour per domain on Netlify, GoDaddy, Vercel, Squarespace, or Wix, and the price for most small businesses is $0.
How AI Agents Coordinate And How They Remember: The Two Architecture Choices Behind Every Agent Project
Two questions decide whether an AI agent project ships or stalls: how the agents coordinate, and how they remember. A small-business and local-government playbook.
Why I Built Site Migration Gen. The Rebuild That Lost a Client Their Brand, Their Rankings, and a Quarter of Their Traffic
I rebuilt a small-business site with a generic AI prompt and shipped a faster, cleaner version that quietly destroyed the brand. This tool is the prompt I wish I had written that day.
Your JSON-LD Passes Validation And Still Breaks — The Graph Problem
Generic JSON-LD validators check each node in isolation and pass your schema with a green checkmark. Then Google ignores half of it because the Article references an author @id that doesn't resolve, or the Organization has no inbound references from anywhere, or the author has no inverse creativeWork link. The linter treats JSON-LD as a graph.
When Every Link To /Pricing Says 'Pricing' — Internal Anchor Over-Optimization
External anchor-text over-optimization is a known backlink-manipulation signal. Internal anchor-text over-optimization is the same pattern on your own site — every link to /pricing saying 'pricing' with zero variation. Shannon entropy measures it. Under 1.5 is a red flag. The scanner crawls your sitemap and measures it per target URL.
Robots Meet AI Employees: A 4-Year Industry Roadmap Based on Real Dev Speed (2027, 2028, 2029, 2030)
Cited 1/2/3/4-year forecast for humanoid robots and AI employees, grounded in actual dev speed. Figure AI, Tesla Optimus, Apptronik, Agility, NVIDIA GR00T, Physical Intelligence — what shipped, what's planned, what banks are projecting.
When the Robot Cooks, Drives, and Runs Errands: Personal Services Automation for Restaurants, Delivery, Trades, and Pilots — Cited Roadmap Through 2030
Cited roadmap for consumer-facing physical AI — restaurants (Miso, Chipotle), delivery (Starship, Nuro), driving (Waymo, Tesla FSD, Zoox), pilot/aviation autonomy (Joby, Wisk), and skilled trades robotics. What's deployed in 2026, what ships through 2030.
Google's AI Overviews Now Pulls Expert Advice From Reddit and Forums. What That Means For Your Content Strategy
Google updated AI Overviews to include preview perspectives from Reddit, web forums, and other firsthand sources. Here is what changes for your SEO, your content, and which audits to run now.
AI's Wider Wave: Legal, Medical, Hospitality, Housekeeping, and Government — State-by-State Impact and a W-2 Survival Playbook
Cited deep-dive on AI in legal, medical, hospitality, housekeeping, and government work. Brookings + Goldman + OpenAI research on which US states are most exposed. A practical W-2 navigation playbook for the next four years.
AI Employees for Small Business: 10 Real Stacks That Move Faster — Cited Cases from Anthropic, OpenAI, Shopify, GitHub, and the Indie Founder Underground
Cited examples of small businesses and solo operators using Anthropic, OpenAI, Shopify Magic, GitHub Copilot, and indie tools to multiply output without hiring. Stacks, costs, and the productivity numbers that vendors actually publish.
AI Employees in 2026: What They Already Do Across HR, Front Desk, SEO, Real Estate, and More — Plus What's Shipping in 2027
Cited capabilities of AI agents across HR, reception, SEO, real estate, sales, customer service, legal, healthcare, software, and marketing — with what shipped in 2026 and what's on the 2027 roadmap.
The .well-known Directory: Your Site's Machine-Readable Identity Hub
The .well-known directory is how machines discover your site's identity, security contacts, AI capabilities, and federation endpoints. Here's the complete guide to every file that matters in 2026.
The Tribute Microsite: When robots.txt Points at Another Domain on Purpose
A focused sub-domain microsite whose robots.txt and sitemap.xml point back to a master domain looks like a misconfiguration. It can be a deliberate authority funnel for campaigns, milestones, and launches. Here's when it works.
Site Migration Gen — Rebuild Your Site Without Losing What Matters
Paste your old site's URL, pick a modern framework and host, and generate a comprehensive AI prompt that rebuilds your site with every best practice baked in while preserving your business name, photos, content, and search rankings.
Migration Analyzer — Know What Breaks Before You Move Your Site
Before migrating your site to a new platform, run the Migration Analyzer. It detects your current CMS, catalogs portable vs vendor-locked assets, scores migration risk, and generates an AI prompt for the rebuild.
A German Page With USD Prices Is Not Actually Localized
Hreflang tags are one locale signal. HTML lang, og:locale, schema priceCurrency, addressCountry, telephone country code, and visible price symbol are five more. They need to agree. When they don't — German page declaring EUR in schema but showing USD prices in the body — Google treats the page as half-localized and conversions drop 20-40%.
Image Prompt Gen — AI Image Prompts for Favicons, Heroes, and Brand Photos
Generate platform-ready prompts for Midjourney or Ideogram to create favicons, hero images, social cards, and brand photos. Includes ImageMagick post-processing commands for every image type.
Apple Says 30 Characters, Google Says 50 — Per-Locale ASO Parity
Apple caps iOS titles at 30 characters, subtitles at 30, keyword field at 100 (and penalizes duplicates from the title). Google Play caps titles at 50, short descriptions at 80. Per-locale. Most apps ship the English listing as a literal-word-count-estimate translation and get truncated or penalized in 15 locales.
I shipped a tool that probes your site as ChatGPT, Claude, Perplexity, and 13 other bots so you can see exactly who is being blocked
The Bot Allowlist Validator sends a real HTTP request to your URL using each AI bot's published User-Agent and reports ALLOWED, WAF-CHALLENGED, BLOCKED, or PARTIAL per bot. Here is why it exists and how to read the output.
When 200 OK Lies: The SPA Shell Trap That Hides /llms.txt and /.well-known/ From AI Crawlers
If your SPA catch-all sends every unknown path to /index.html with a 200, AI crawlers think you have an llms.txt, a security.txt, and an agent-card.json. They don't. Here's how to detect it, and the Netlify config that actually fixes it.
Netlify Database, in Practice: Why a Postgres Branch per Deploy Preview Changes the Workflow
Netlify Database pairs Postgres with Deploy Previews so every PR gets an isolated database branch with a copy of production data. Here's what that lets you actually do, with code examples for migrations, schema changes, and AI-agent-driven feature builds.
How Mattel Consolidated 500 Digital Properties on One Platform — and What It Teaches Solo Operators
Mattel went from 500+ scattered web properties on legacy infrastructure to a single platform their full-stack developers can own end-to-end. The same principles work at one site or fifty. Here's what to actually take from the case.
ImageMagick for Favicon Sets and Brand Image Processing
Complete ImageMagick command reference for generating multi-size favicon.ico files, apple-touch-icons, PWA icon sets, WebP hero images, and OG social cards from a single source image.
Your Site Has No Favicon and It's Costing You Trust
A missing favicon makes your site look unfinished in browser tabs, bookmarks, and search results. Here's what to add, what sizes matter, and why AI engines use icon presence as a trust signal.
The Zero-Request Favicon: Emoji-as-data-URI-SVG You Can Ship in 60 Seconds
Skip the 16x16, 32x32, 180x180 favicon set entirely. One line of HTML, an emoji, and an SVG wrapper get you a sharp scalable favicon with zero network requests. Here's the recipe and the trade-offs to know.
Three Signals That Tell AI Engines Your Site Is Abandoned, Unqualified, or Locked Behind a Vendor
The Mega Analyzer's cross-validation layer detects abandoned tech, YMYL gaps and vendor-locked platforms — and now lets you correct a misread vertical or platform so every check and fix prompt regenerates for your real positioning.
Cloudflare is blocking AI crawlers from your site. Here's how to fix it without weakening bot protection.
If your site sits behind Cloudflare bot management, ChatGPT and Perplexity see a 403 challenge page with noindex/nofollow. Here is what is happening and how to fix it.
How to Generate Favicons and Brand Images with Midjourney and Ideogram
Step-by-step walkthrough of generating favicons, hero images, and brand photos using Midjourney's Discord bot and Ideogram's web UI plus API. Includes prompt formulas, API code, and ImageMagick post-processing.
Apple, Spotify, And Podcasting 2.0 — One RSS Feed, Three Validation Passes
Podcasting 2.0 adds transcript, chapter, person, and value tags that power AI-answer citations for audio content. Most feeds ship without them because Apple's submission flow doesn't require them. The new audit validates Apple basics, Spotify requirements, and Podcasting 2.0 namespace in one pass.
How One E-Commerce Category Page Spawns 10,000 Crawl-Waste URLs
Param-crawl-waste counts single-parameter duplicates. Facet-trap does different math: on a category page with 5 facets averaging 4 values each, the combinatorial space is 5 × 5 × 5 × 5 × 5 = 3,125 URLs. For 12 products. Google wastes 99% of its crawl budget on that one category.
Why Your YouTube Video Isn't Getting Key Moments Chapters In Google
Google's Key Moments SERP carousel only fires when your chapter formatting passes strict rules: first timestamp at 0:00, at least 3 chapters, 10-second minimums, ascending order. Or your page has proper Clip / SeekToAction JSON-LD. The new checker validates both paths.
When Programmatic SEO Becomes Thin-Content Spam — Shingle Detection
Programmatic SEO scales content by swapping one noun across a template. Google's Helpful Content system detects near-duplicate sets via shingle similarity and filters the whole template when pairwise similarity stays above 0.7. The new detector runs the same math in your browser.
The 17 Google Merchant Listings Fields Generic Validators Miss
Generic JSON-LD validators green-light Product schema that Google's Shopping Graph rejects. The Merchant-specific subset — shippingDetails, hasMerchantReturnPolicy, energyEfficiencyCategory, pricePerUnit, global identifiers — has its own rulebook, and as of 2026 Google enforces it more strictly for EU listings.
INP Attribution — Which Click Actually Broke The 200ms Budget
CWV Audit says your INP p75 is 340 ms. That's the aggregate. It doesn't tell you which click, which handler, which third-party script caused the jank. The new profiler loads your URL in an iframe, hooks PerformanceObserver + long-animation-frame, and attributes each slow interaction to a specific event target and script origin.
Pick Your Web Stack Without Guessing — A Visual Guide to 85+ Free Tools for Small Business Owners
You have a website. You know it could be better. You have no idea which of the 85+ free tools on jwatte.com apply to your situation. This guide groups every generator and audit tool by how much work it is, what it protects, and when a small-business owner should care — with concrete paths for three common scenarios.
GPTBot Doesn't Run JavaScript — What Your Site Looks Like To It
GPTBot, ClaudeBot, PerplexityBot, and CCBot don't execute JavaScript. They read your raw server HTML and move on. If your content only exists in the JS-rendered DOM, they see an empty shell. The audit catches thin raw HTML, render-blocking scripts, head-heavy CSS, and missing title/meta in the server response.
When 80% of Your Crawl Budget Goes to ?sort=price Duplicates
E-commerce and large blog sites routinely spray URLs like /products?sort=price&filter=red&page=2&utm_source=email. Googlebot crawls every combination; 80% of crawl budget goes to near-duplicates of one canonical page. The auditor clusters URLs by parameter fingerprint and emits the robots.txt + canonical config to stop the waste.
IaC Deploy Generator — One Form, Terraform + Ansible, Five Clouds, Production-Hardened
Pick a cloud, a reverse proxy, and the services you want running. Emits Terraform HCL for provisioning (VM + firewall + floating IP) plus an Ansible playbook for configuring (Docker, Caddy/nginx/Traefik, TLS, the services). A repeatable, idempotent path from zero to HTTPS-serving host in 15 minutes.
The same person built RSS, RDF, Schema.org, and NLWeb. That is mostly the whole story.
R.V. Guha has spent thirty years giving the web a meaning layer for whatever client comes next. NLWeb is the latest pass at the same idea, not a new direction.
The agent web has four protocols now. Here is where each one fits.
NLWeb, WebMCP, ACP, UCP. Four open standards stack rather than compete. Here is what each one actually does, and which signals you can ship today.
7 Ways to Build a Local Business Lead List Without Buying One
A practical comparison of lead generation methods for local businesses: Google Places API, Yelp Fusion, OpenStreetMap, state registrations, industry directories, SerpAPI, and Apollo.io. Costs, limits, and code.
You Got a California Tracking Demand Letter. Here's How to Find Out What's Actually on Your Site.
Step-by-step guide to auditing your website after receiving a CCPA or pixel-based tracking demand letter. Free tools to identify every third-party script, cookie, and consent gap on your site.
No viewport meta tag means Google indexes your site as desktop-only
Without a properly configured viewport meta tag, mobile browsers render your page at 980px wide and shrink it to fit. Google's mobile-first indexing treats this as a non-mobile page.
Your video has 10,000 views and zero search impressions because the schema is missing
VideoObject schema markup tells search engines what your video is about, how long it is, and where the thumbnail lives. Without it, your video pages are invisible in video search results.
Your site doesn't respect dark mode. Your visitors notice.
CSS media queries for prefers-color-scheme, prefers-reduced-motion, and prefers-contrast let your site adapt to user preferences. Most sites ignore them.
Your URLs are telling search engines more than you think
Inconsistent trailing slashes, uppercase letters, session IDs in paths, and deep nesting all cost you crawl budget and split link equity. A URL hygiene audit catches them.
Trusted Types block the DOM XSS attacks that CSP alone cannot stop
Content Security Policy blocks inline script injection but does nothing about innerHTML and eval sinks. Trusted Types close that gap at the browser level.
That one marketing tag is costing you 600ms and you don't even check the dashboard
Every third-party script added to your site costs main-thread time, delays interactivity, and ships visitor data to someone else's server. Most sites have tags nobody uses anymore.
If your text contrast fails WCAG, one in twelve male visitors can't read it
Low-contrast text isn't a design preference. It's a readability failure that affects 300 million people with color vision deficiency and everyone reading in sunlight.
HTML tables without headers are invisible to screen readers
A table without th elements, scope attributes, or a caption is just a visual grid. Screen readers can't announce what column a cell belongs to, so the data becomes meaningless.
Your tab order is broken and keyboard users already left
A logical tab order is a WCAG 2.1 requirement and a usability baseline. If pressing Tab skips fields, jumps to the footer, or traps focus in a modal, keyboard-only visitors can't use your site.
Structured data is not optional anymore and most of yours is broken
Google uses structured data to generate rich results, knowledge panels, and AI overviews. Missing or malformed schema means your content gets displayed as a plain blue link.
One missing integrity hash and your CDN becomes an attack vector
Subresource Integrity ensures external scripts and stylesheets have not been tampered with. Without it, a compromised CDN can inject code into every page load.
Speculation Rules will make your site feel instant if you set them up right
The Speculation Rules API lets browsers prerender pages before the user clicks. When configured correctly, navigations feel instant. When misconfigured, you waste bandwidth.
Your reviews exist but Google cannot see them
Star ratings and review counts on your page mean nothing to search engines without proper schema markup. Most sites display social proof visually but fail to mark it up for rich results.
Why auditing one page at a time misses the pattern
Single-page audits catch page-level issues. Site-wide sampling catches template-level issues that affect hundreds of pages at once. Here is the difference.
One giant sitemap is a crawl budget problem hiding in plain sight
A single sitemap.xml with thousands of mixed URL types makes it harder for search engines to prioritize your important pages. Segmented sitemaps fix that.
Your service worker might be caching problems instead of solving them
A misconfigured service worker can serve stale pages, break updates, and trap users in an old version of your site. Here is what to check and why it matters.
Your service worker might be caching pages it should not touch
A misconfigured service worker scope can serve stale content, break navigation, or cache private pages. Most sites never audit what their service worker actually controls.
You might not even be in the race and here is how to check
Before you optimize a single page, find out whether Google considers you a competitor for your target queries. Most small sites are not even in the cohort.
You found 200 SEO issues. Now what? Build a roadmap that doesn't stall.
SEO audits find problems. Roadmaps fix them. A 30/60/90 day plan with impact-effort scoring turns overwhelming findings into a sequence you can actually execute.
Your robots.txt is probably blocking something you did not intend
A single misplaced wildcard in robots.txt can hide entire sections of your site from search engines. Most site owners never test their rules against real URLs.
You have schema markup. Google still won't give you a rich result. Here's why.
Partial schema triggers validation but not rich results. Missing required fields, wrong nesting, and incomplete types are why your snippets don't appear.
You're sending a 2400px image to a 375px phone. Here's what that costs.
Responsive images with srcset and sizes serve the right resolution to each device. Without them, mobile users download images 5-10x larger than needed.
Your homepage is 8MB of HTML. Nobody asked for that.
Oversized HTML responses waste bandwidth, slow parsing, and hurt mobile users. A response size budget catches bloat before it costs you traffic.
Preload, prefetch, preconnect: you are probably using the wrong one
Resource hints tell the browser what to load and when. Using preload where you meant prefetch wastes bandwidth. Using prefetch where you meant preload delays rendering.
Your site is breaking in production and nobody is reporting it to you
The Reporting API lets browsers tell you about CSP violations, network errors, and deprecation warnings automatically. Most sites never set it up.
Why your site's invisible supply chain matters more than you think
Every externally loaded font, script, and tracking pixel is a dependency you don't control. A single CDN outage or privacy policy change can break your site or expose your visitors.
Your outbound links are making promises you might not intend
Every link on your page tells search engines something about trust. A missing rel attribute on a sponsored link or user comment can cost you ranking authority.
When alt text repeats the heading, screen readers say everything twice
Images with alt text that duplicates the adjacent heading or caption create a stuttering experience for screen reader users. Here is how to find and fix every instance.
Every redirect hop costs you speed, PageRank, and crawl budget
Redirect chains of 3+ hops waste crawl budget, dilute link equity, and add hundreds of milliseconds to every page load. Most sites have chains they do not know about.
Your content reads at a college level but your audience didn't go to college
The average American reads at a 7th-8th grade level. If your site content reads at grade 12+, you are writing for yourself, not your audience. Here is how to measure it.
If Google can't see your content without JavaScript, neither can AI
SPA frameworks that render everything client-side ship empty HTML to crawlers. Your content exists in JavaScript bundles that search engines may never execute.
Broken pagination markup quietly deindexes your best content
When rel=next/prev is wrong or your canonical points every paginated page back to page 1, Google may ignore your deep archive entirely. Here is how to check.
Pages in Your Sitemap With Zero Internal Links — Orphan Detection
A page in your sitemap that no internal link reaches from the homepage is an orphan. Googlebot discovers pages primarily through links — sitemap-only pages get deprioritized or never crawled. The detector BFS-crawls your site from the homepage up to N hops, then diffs against your sitemap.
Your pages look broken on social media because nobody checked the meta tags
Missing or misconfigured Open Graph and Twitter Card tags mean your content shows up on social platforms with wrong images, truncated titles, or blank previews.
The CSS animations killing your INP score
Animations on color, box-shadow, width, and top trigger paint on every frame. The browser works harder, the main thread blocks, and your INP suffers.
When your meta robots and X-Robots-Tag disagree, pages vanish
A noindex in your HTTP headers silently overrides what your HTML says. Most site owners never check both layers, and pages disappear from search without warning.
If a page is six clicks from your homepage, search engines might never find it
Click-depth directly affects how often Googlebot crawls your pages and how much PageRank they inherit. Most sites bury their best content too deep without realizing it.
Mixed content is still breaking sites in 2026
One http:// image on an https:// page can trigger browser warnings, break trust indicators, and get your page flagged. Here is how to find every instance.
When meta robots and X-Robots-Tag disagree, your pages silently vanish
A noindex in your HTTP headers can override everything in your HTML head. Most site owners never check both layers, and their pages disappear from search without warning.
llms.txt is the new robots.txt and most sites are doing it wrong
AI answer engines need structured guidance about your site. A broken or missing llms.txt means LLMs guess instead of citing your best content.
Your outbound links tell search engines who you trust
Every link you place on a page is a vote of confidence. Link to junk and search engines notice. Link to authoritative sources and your own credibility rises.
Lazy loading every image is not a performance strategy
Adding loading=lazy to every image sounds like a quick performance win, but it delays your LCP element and creates layout shifts. An audit separates what should be lazy from what should not.
Your JSON-LD validates but Google ignores it because the graph is broken
Valid JSON-LD syntax doesn't mean a working schema graph. Dangling @id references, orphan nodes, and broken nesting silently prevent rich results.
That popup is costing you rankings. Google told you in 2017.
Google penalizes intrusive interstitials on mobile. Full-screen popups, forced newsletter modals, and cookie walls can suppress your search visibility.
Client-side validation is a courtesy, not a guardrail
Auditing client-side input validation to catch missing constraints, inconsistent error messages, and patterns that break on mobile or with autofill.
Every onclick= in your HTML is a security hole waiting to happen.
Inline event handlers block Content Security Policy enforcement and create XSS vectors. Here's how to find them and refactor to addEventListener.
Your images aren't in Google Images because they're not in your sitemap.
Standard sitemaps list pages, not images. The image sitemap extension tells Google exactly which images to index and what they depict.
Your hero image is lazy-loading. That's why your LCP score is terrible.
Lazy-loading above-the-fold images delays your largest contentful paint. Here's how to audit LCP candidates and fix the loading priority chain.
The missing width and height attributes silently tanking your CLS score
Images without explicit width and height attributes cause layout shifts that hurt Core Web Vitals. The fix is two HTML attributes per image tag.
When five links all say 'Click here,' screen readers hear five identical doors to nowhere
WCAG 2.4.4 requires link text to describe the destination. When multiple links share the same visible text but point to different URLs, keyboard and screen reader users can't tell them apart.
Your how-to content does the work but Google can't read the steps
HowTo schema markup turns step-by-step content into rich results with expandable steps. Most implementations miss required fields or break nesting rules.
Most forms fail accessibility before anyone ever fills them out
Unlabeled inputs, missing autocomplete hints, and broken tab order silently exclude users and invite ADA demand letters. Here is what the audit checks and how to fix it.
You're shipping 200KB of font glyphs your visitors will never see
Most web fonts include thousands of glyphs for languages your site doesn't support. Subsetting strips the dead weight and cuts font payload by 70-90%.
Google Fonts adds 400ms to your first paint. Self-hosting the same files takes 10 minutes.
Loading fonts from fonts.googleapis.com requires two extra origins, blocks rendering, and leaks visitor IPs to Google. Self-hosting the same WOFF2 files eliminates all three problems.
Why your site feels slow even when PageSpeed says it's fast
First Input Delay measures the gap between a user's first click and the browser's response. A page that loads fast but responds slowly still feels broken.
Featured snippets follow a pattern. If your content doesn't match it, you won't get one.
Google pulls featured snippets from pages with question-shaped headings followed by concise 40-60 word answers. Most pages bury their answers in paragraphs no algorithm can extract.
Your FAQ schema probably doesn't match what's actually on the page
Ghost questions in your JSON-LD, missing answers, and drift between visible FAQ content and schema markup cost you rich results silently.
Your events page is invisible to Google if the schema is wrong
Event schema errors silently kill your chances at rich results. Missing dates, wrong time zones, and stale listings are the most common culprits.
You're paying for a CDN. Your cache hit ratio says it's not working.
Most sites push traffic through a CDN and assume it's caching. Response headers tell the real story. Cache-Status, Age, and X-Cache reveal whether your edge is serving or just proxying.
Your server knows what the browser needs before the HTML arrives. Early Hints tells it.
HTTP 103 Early Hints lets your server push preload and preconnect directives while it's still building the response. Most servers support it. Almost nobody uses it.
DNS Records Generator — One Email Setup, Five Provider Syntaxes, Zero Typos
Email authentication records that pass Gmail delivery tests on the first try. Pick an email provider (Google Workspace, Microsoft 365, Fastmail, Mailgun, SES, etc.), enter your domain, and get the full SPF + DKIM + DMARC + MX + CAA + MTA-STS + DNSSEC record set in whichever provider syntax your DNS host speaks.
Your CSS is probably three times heavier than it needs to be
CSS bloat slows rendering, increases specificity wars, and makes every design change a guessing game. Here's how to measure complexity before it buries you.
Most CSP headers use unsafe-inline and call it security. Here's what strict actually looks like.
A CSP with unsafe-inline stops nothing. Nonce-based policies, strict-dynamic, and Trusted Types are the three layers that actually prevent XSS. Most sites have zero of them.
Your CSP says allow everything from *.cdn.com. That's not a policy, that's a suggestion.
A Content Security Policy full of wildcards protects nothing. The CSP Allowlist Completeness Audit cross-references what your page actually loads against what your headers permit.
Your Lighthouse score is lying to you. Here's what Google actually measures.
Lighthouse runs on your machine with your fast CPU and your empty cache. CrUX field data is what Google ranks on. The gap between the two costs real traffic.
You can't beat competitors you haven't measured
A side-by-side SEO comparison across six dimensions reveals where you're ahead, where you're behind, and where the gap is small enough to close.
The User-Agent string is dying and your site probably isn't ready
Chrome is freezing the User-Agent string and replacing it with Client Hints. If your site relies on UA parsing for device detection, it will break.
If AI can't find you in its training data, you don't exist to its users
AI answer engines cite brands they can verify across Wikipedia, Wikidata, G2, Crossref, and 14 other corpora. If you're absent from these, you're invisible.
RAG Splits Your Page Into Chunks — How To Make Each Chunk Retrievable
Retrieval-Augmented Generation splits your page into fixed-token chunks, embeds each, and retrieves the best-matching chunk when a user asks a question. If your content only makes sense when read in order, RAG fails. The scorer splits your URL three ways and grades each chunk.
Your breadcrumbs are lying to Google about how deep your site goes
Breadcrumb trails that skip levels, dead-end on category pages, or nest seven layers deep confuse both crawlers and humans. Here's how to audit the depth.
Google checks whether your articles have a real author and most don't pass
E-E-A-T starts with authorship. Eight signals determine whether Google treats your article as credible or anonymous content. Most sites miss at least four.
Your API keys are in your HTML and anyone can read them
Leaked API keys in frontend code cost businesses thousands in unauthorized charges. Most developers don't realize their secrets are visible in page source.
AI wrote your content and everyone can tell except you
Google, the EU AI Act, and readers all expect AI-generated content to be disclosed. Most sites miss the signals. Here is what to check and how to fix it.
Your 404 page is a ranking signal and most of them are terrible
A bad 404 page loses visitors and tells search engines you don't care about user experience. Here's what Google actually checks and how to fix it.
How to Build a Self-Healing Lead Pipeline with the Google Places API
Step-by-step build of an automated lead generation pipeline using Claude Code, Google Places API, Node.js, and Instantly.ai. Includes the prompt, the code, and the full import walkthrough.
ZeroSSL — The Let's Encrypt Alternative I Reach for When Rate Limits, Wildcards, or IoT Get in the Way
ZeroSSL is an ACME-compatible certificate authority that solves three specific problems Let's Encrypt doesn't: its 90-day rate-limit behavior on bulk issuance, wildcard issuance without DNS-01 on every provider, and 90-day certs for a large fleet of IoT devices. Here's when it actually matters.
XPU, NPU, GPU, TPU: What Each AI Chip Actually Does and When You Need Which
NVIDIA dominates AI training. But for inference, edge AI, and cost-sensitive workloads, NPUs, TPUs, and Intel's Gaudi are closing the gap. Here's the 2026 hardware map with actual TOPS numbers, cost comparisons, and the use case each chip type was built for.
Every Redirect Hop Beyond the First Leaks PageRank — Bulk Chain Audit
Paste up to 200 URLs. The auditor follows every chain, reports final status, flags chains longer than 2 hops (PageRank leak), catches mixed http → https downgrades, detects meta-refresh chains, and flags canonical mismatch at the final URL. Same signal Screaming Frog reports, free + browser-only.
Multi-Location Schema Audit: Why Franchise Sites Get Structured Data Wrong
Businesses with 2+ locations almost always have schema problems. One LocalBusiness for five branches, missing geo coordinates, NAP drift between pages.
Why a Multi-Location Schema Audit Exists
Franchise and chain businesses often ship generic LocalBusiness schema for every location. This audit validates subtypes, geo coordinates, and parent-child relationships per site.
Generate a Spec-Compliant /llms.txt From Your Sitemap in One Click
Point the generator at your sitemap. It fetches each URL's title + meta description, groups by path pattern into H2 sections, and emits a spec-compliant /llms.txt with H1 title, blockquote description, and link-list formatting. Diffs against any existing /llms.txt.
IndexNow Submission Audit: Stop Waiting for Search Engines to Find Your Pages
Your IndexNow key file might be broken and you wouldn't know it. Here's what the protocol requires, what goes wrong, and how to audit your setup in 30 seconds.
Duplicate Title and Meta Description Audit: The SEO Problem Hiding in Plain Sight
Duplicate titles and meta descriptions confuse search engines and waste crawl budget. Here's how they happen, why they matter, and how to find every instance.
Why a Duplicate Title and Meta Description Audit Exists
Duplicate titles dilute SERP click-through and duplicate descriptions get replaced by Google's own snippets. This tool crawls your sitemap to find every match.
Where Claude Opus 4.7 Actually Ranks and What Early Adopters Learned
Opus 4.7 leads SWE-bench at 87.6% and Chatbot Arena at 1504 ELO. It also lost 46 points on long-context retrieval and costs 34% more tokens for the same prompts. Here's where it actually stands, benchmark by benchmark, and what the first wave of upgraders paid.
Mercor: The $10 Billion AI Hiring Platform, Its Data, and the Cracks Showing
Mercor hit a $10 billion valuation at age two, manages 30,000 contractors paid $1.5M per day, and just confirmed a data breach from a supply-chain attack. Here's what the platform does, how the money flows, what the data risks look like, and why Scale AI is suing them.
A Markdown File Is the Best Memory Layer for Your AI Coding Tool
Three of the biggest AI coding platforms chose plain Markdown files over vector databases for persistent memory. Here's why it works, how to set it up in Claude Code, Codex, Cursor, and others, and when you actually need something heavier.
Search Console + Bing Importer: Cross-Reference Your Coverage Data Before It Costs You Traffic
GSC and Bing Webmaster Tools show coverage data in separate silos. This tool merges both CSVs against your sitemap to find the pages nobody is indexing.
DeepSeek V4 vs Kimi K2.6 vs Claude vs GPT: Where Open-Weight Models Actually Stand
DeepSeek V4 Pro hits 80.6% on SWE-bench Verified at $0.30 per million tokens. Kimi K2.6 scores 80.2% at $0.60. Both are open-weight. Here's how they compare to Claude Opus 4.7 and GPT-5.4 on coding, cost, and real-world tasks.
Think Like McKinsey For Free: Consulting Frameworks You Can Run Through Claude or Codex
McKinsey charges $500K per engagement. The frameworks they use are public knowledge. MECE, the Pyramid Principle, 7S, Porter's Five Forces, the BCG Matrix. Here's how to run each one through an AI coding CLI and get 80% of the analysis at 0% of the cost.
BYD Electric Cars: The Privacy Problem Nobody Reads Before Signing
BYD's God's Eye ADAS system includes 12 cameras, 5 radars, and 12 ultrasonic sensors as standard. It also sends data to Chinese servers via a GSM modem. Here's what the tech does well, what the privacy risks look like, and how BYD compares to Tesla on both fronts.
Build an AI Job Search Agent and an Inference Engine That Sizes Itself
The job search is a pipeline problem. Here's how to build an AI agent that finds openings, tailors resumes, and applies automatically, plus an inference routing layer that sends simple tasks to cheap models and complex ones to expensive models.
Babel Audio Pays Up To $50 an Hour To Read Aloud — But It Has To Be Perfect
David AI Labs runs Babel Audio, a speech data platform that pays up to $50 per hour for recorded conversations and read-aloud tasks. The catch: perfection is the baseline. Here's how it works, what the pay actually looks like, and whether it's worth your time.
Five AI Agents That Replace Five Hires: How To Build a Workflow Staff With Claude or Codex
An administrative assistant, a researcher, a chief of staff, a life coach, and a finance analyst. Five agent roles you can configure in Claude Code or Codex today, with the system prompts and workflow patterns that make each one usable.
Every Time an AI Model Update Broke Something
GPT-4's prime-number accuracy dropped from 97.6% to 2.4% in three months. Claude's thinking length collapsed 73%. Opus 4.7 lost 46 points on context retrieval. Here's the full timeline and what each incident teaches about validating before you upgrade.
GPT-5.5 (Spud) vs Claude Opus 4.7 — What Actually Changed At The CLI, And Why The Lead Isn't Durable
OpenAI shipped GPT-5.5 (codename Spud) and Anthropic countered with Opus 4.7 inside a month. Benchmarks move; the experience at the terminal shifts less than the marketing suggests. Here's what actually changed, what it means for Codex CLI vs Claude Code users, and the pricing squeeze Michael Parekh called correctly.
How To Validate an AI Coding Model Before You Trust It With Your Codebase
AI coding tools ship model updates that can silently break workflows, spike token costs, or claim tasks are done when they aren't. Here's a concrete pre-upgrade checklist and six alternative CLIs to keep warm.
SGE Readiness Audit: 18 Signals That Determine Whether Google AI Overviews Cite Your Page
A free tool that scores any URL against 18 on-page signals correlated with inclusion in Google AI Overviews. Checks answer conciseness, schema depth, passage retrievability, and E-E-A-T markers.
Why Scroll Depth + CTA Visibility Map Exists
Maps every CTA element to its scroll-depth position on mobile and desktop, flags buttons buried past 75% scroll, and identifies long dead zones with no visible action.
Why Revenue-Per-URL Attribution Map Exists
Paste a GA4 Pages-and-screens export and see which 50 pages earn 80% of your revenue, classified as Stars, Hidden Gems, Vanity Rankers, or Cut Candidates.
Why Pricing Psychology Audit Exists
Scans a pricing page for twelve psychology levers that move purchase decisions: tier count, decoy structure, anchor strikethrough, annual framing, guarantee placement, and more.
What Topics Your Competitors Cover That You Don't — Heading Gap Audit
Enter your article URL + 2-5 competitor URLs ranking for the same query. The tool extracts every H2 and H3 from each, builds the union of topics, shows what competitors cover that you don't (and vice versa), and scores your topical coverage %. Free alternative to Clearscope and Surfer.
Why Form-Field Friction Estimator Exists
Every form field has a drop-off cost. The estimator classifies each field by type, applies Zuko/Formisimo benchmark weights, and predicts your cumulative completion rate.
Why Checkout Abandonment Autopsy Exists
70% of US carts abandon. Half of that is fixable on-page. The audit finds the ten friction patterns responsible and ranks them by expected recovery.
Predicting Which Keywords Fire AI Overviews Without a Paid SERP API
Google AI Overviews fire more often on informational, how-to, comparison, and 'best/top' queries than on navigational or transactional ones. Paid SERP-feature trackers cost $99-$399/mo to report this. The new heuristic scores each keyword in your list from the query structure alone — no SERP API.
Why About-Page Conversion Audit Exists
Your About page is the second-highest-intent page on your site. The audit scores 14 trust and conversion signals most SMB sites miss entirely.
How to Get Multiple Pages in a Single Google Search Result
ItemList carousel schema lets you occupy a horizontal scrolling carousel in Google search — showing multiple pages from your site in a single result. Here's the implementation guide with real CTR data.
How to Send Email That Actually Gets Delivered: A Small-Business Infrastructure Playbook
Set up email so your password resets, newsletters, and outreach all land in the inbox instead of spam. Secondary domain, Google Workspace mailboxes, full SPF/DKIM/DMARC on your DNS host, platform pick by use case, and a 21-day warmup. Updated for Gmail and Yahoo's 2024 bulk-sender rules.
Caddy — The Reverse Proxy That Handles Its Own Certificates (and Quietly Solves Half Your On-Prem TLS Problems)
Caddy gets TLS right by default: automatic Let's Encrypt for public sites, a built-in local CA for internal services, and ACME support for enterprise CAs. I run Caddy in front of public sites, in front of homelab containers, and as the trust anchor for a 40-device internal network. Here's where it fits and where it doesn't.
Qwen — When It's The Right Model And How To Run It At Home On Consumer Hardware
Qwen is the strongest open-weights model family most developers have never run. This covers when Qwen beats Claude / GPT (privacy, cost, offline, fine-tuning), the hardware tiers that matter, a concrete Ollama setup path, and the honest limits compared to frontier proprietary models.
Gemma — Google's Open-Weights Family, And Where It Fits Next To Qwen
Gemma is Google's open-weights sibling to Gemini. Same architectures-in-spirit, different philosophy and release cadence than Qwen. When to pick Gemma over Qwen, how to run the current generation at home, and what the 3 / 4 generation actually changed.
Google Gemini And Its CLI — Long Context, Multimodal, Google-Stack Native. When Each Strength Actually Pays Off.
Gemini's superpower is context length — 1M+ tokens on the frontier variants. That changes which tasks it wins at. This walks through when Gemini is the right choice over Claude or GPT, how the Gemini CLI works in practice, and where Google's ecosystem integrations pay off.
Character Counts Lie — Why Your SERP Snippet Truncates Differently Than You Expect
Google doesn't truncate titles at 60 characters. It truncates at ~580 pixels desktop, ~460 mobile. A title full of 'W's hits the pixel wall 15 characters earlier than one full of 'i's. The new tool measures actual pixels and previews both devices.
WCAG 1.1.1 Plus CLS Risk In One Pass — Image Alt Audit
Every image on a page in one table: alt present, missing, empty, filename-as-alt flagged, length in the 5-125 char band, duplicate alts across images, and missing width/height attributes that introduce CLS. WCAG 1.1.1 compliance + Core Web Vitals CLS risk in one pass.
The Hidden FAQ-Schema Bug Killing Your Rich Results (and the Tool That Finds It)
Google will silently suppress your entire FAQPage rich result if the JSON-LD lists questions that are not visible on the page, or if visible FAQs never make it into schema. Most site owners do not know the rule exists. The FAQ Schema Parity Audit catches both cases in ten seconds and hands you a drop-in JSON-LD block. Here is why I built it, when to run it, and what to fix first.
CWV Fix Generator — LCP, CLS, INP Patches From a Core Web Vitals Audit
The CWV audit tells you LCP is 3.2 seconds. This tool tells you what to paste into your HTML, CSS, and JS to drop it under 2.5. Twenty concrete patterns covering every common LCP, CLS, INP, and TTFB failure, bundled into an AI prompt that tailors them to your page.
Two CLIs, One Workflow — Running Codex Alongside Claude Code Without Losing Your Mind
A split-screen setup that uses each CLI for the tasks it handles best. Claude Code in the main terminal for long work, Codex in a second terminal for fast questions and focused tasks. Here's the practical routine — keybindings, terminal layout, and the rules that keep them from colliding.
Prompt Style — The Real Differences Between GPT And Claude, With Examples That Actually Show It
Both families respond to good prompting. They don't respond to the same good prompting. Four concrete differences, with before-and-after examples showing a prompt that works for one family and needs adjustment for the other. Written for people running both.
Codex vs Claude Code — Task By Task, Where Each One Actually Wins
Not a feature-matrix wall of checkmarks. A task-level breakdown of where Codex wins, where Claude Code wins, where they tie, and where the choice depends on your specific situation. Written by someone running both.
Codex CLI After Install — What To Actually Do With OpenAI's Agentic Terminal Once It's Running
The install primer got you Codex on your PATH. This is the follow-on — GPT model selection for different task types, prompt shape that actually works with GPT families, when to keep Codex open alongside Claude Code, and the OpenAI-specific cost and rate-limit mechanics you'll hit in week one.
ChatGPT CLI, Codex CLI, GitHub Copilot CLI — Three Products That Sound The Same And Do Different Things
A kickstart primer installs all three. Two use GPT under the hood, one uses a different model under GitHub. Their jobs barely overlap. Here's the actual positioning — what each one is for, when to reach for which, and the places people confuse them.
WCAG Fix Generator — Turn the Audit's Failure List Into an AI Remediation Prompt
The WCAG audit finds 50 accessibility failures. Manually writing fix instructions for 50 failures is why teams never finish. This tool emits a single AI prompt with per-failure HTML/CSS/ARIA patches + a prioritized 30-day WCAG 2.2 AA + ADA compliance roadmap.
The Pre-Deploy Hook That Kills Regressions Before Netlify Ships Them
A drop-in Node script that runs 15 regression checks against your built _site/ folder before every Netlify deploy. Blocks force=true redirect loops, missing sitemap.xml, template leakage, orphaned pages, stale tool references, broken in-page anchors. Exit 1 = deploy aborted.
AI Answer Engines Rank at the Paragraph Level — Here's How to Tell Which Paragraphs Win
Perplexity, ChatGPT Search, and Gemini rank at the passage level, not the page level. A page with one retrieval-ready paragraph outperforms a page with ten average paragraphs. The scorer splits any URL into passages and scores each on six signals — length, direct-answer opener, pronoun density, fact density, self-containment, subject clarity.
Polarity, Hedging, and Risk Phrases — Three Signals in Any Brand Response
Paid brand-sentiment tools train NLP models on proprietary corpora and sell the result at $99/month. A good lexicon does 80% of the work and runs in your browser. The scanner scores polarity, flags hedge words, and surfaces risk phrases — tuned for brand-mention contexts in AI responses.
CLAUDE.md Anti-Patterns — What Not To Put In, Why Your File Rots, And How To Fix It Before It Hits 800 Lines
Your CLAUDE.md started clean and is now a graveyard of decisions you stopped making, rules you never enforce, and tribal knowledge no one remembers writing. Here are the 8 anti-patterns that cause it — with fixes for each.
Skills, Rules, Memory — Where Each One Actually Belongs (And Why Your CLAUDE.md Keeps Growing)
If every new project fact lands in CLAUDE.md, the file gets to 800 lines and the model stops reading carefully. Here's the four-layer hierarchy that scales — CLAUDE.md, rules, skills, memory — with real examples for each and a migration path when your CLAUDE.md is already a mess.
Remote Control + Channels: Run Your Claude Code Session From Your Phone (iMessage, Telegram, Discord)
Remote Control turns a running Claude Code session into something you can reach from your phone or browser. Channels add chat-app transport so a prompt from iMessage or Telegram lands in your session. Here is the accurate setup, the security model, and when it is worth the permission grants.
Claude Is Not Always The Right Model — When To Break Out Of The Monoculture, And How To Route Between Providers Without Losing Your Mind
Defaulting to Claude for everything is the 2026 version of defaulting to jQuery for everything. The tier-up is routing — Claude for code edits, a cheaper model for bulk classification, a faster model for inline autocomplete, a second opinion when you want one. Here's when and how.
/loop Is A Watchdog, /schedule Is An Alarm Clock — Pick The Right Claude Code Automation For The Job
Both let you automate recurring work, but they solve opposite problems. /loop lives inside an active session and dies when it closes. /schedule runs remotely on a cron and never needs you present. Here's when to use which, with working example tasks for each.
The Weekly /insights Iteration — 30 Minutes Every Monday To Make Claude Code Faster For You
Most people run /insights once, browse the pretty charts, and never open it again. The tier-up is reading it every Monday for 30 minutes and fixing the top friction point. Six weeks of this compounds into a workflow that is measurably faster than when you started.
Stop Approving The Same Prompt 40 Times A Day — A Working Hooks Implementation Guide For Claude Code
You clicked approve on npm test 47 times today. A 12-line hook removes the prompt forever without removing the safety. Concrete PreToolUse, PostToolUse, Stop, and UserPromptSubmit examples you can copy into .claude/hooks.json today.
Auto Mode, In Practice — Four Case Studies Where The Classifier Got It Right, Wrong, And Almost Right
Auto Mode removes approval fatigue for routine actions. That's the pitch. In practice, the classifier is excellent on the boring middle, shaky on two specific edges, and wrong in a way you should understand before turning it on. Four real cases with the fix for each.
The Agent Skill Marketplace — What Skills 2.0 Actually Unlocked, How To Evaluate A Skill Before Installing, And When To Publish Your Own
Skills 2.0 turned custom instructions into programmable agents you can install, compose, and publish. The marketplace is now real — skilz, claude-plugins-official, community repos. Here's how to find skills worth installing, how to audit one before trusting it with your repo, and how to publish yours if you have one worth sharing.
Make The Subagents Fight — Adversarial Review With Worktree Isolation, Boris Cherny's Pattern For Non-Trivial Changes
Five parallel Claudes on five separate git worktrees is the pattern Boris Cherny built Claude Code around. Here's how to run it in your own repo — the review roles that actually catch bugs, the merge strategy that doesn't blow up, and the places the pattern breaks down.
Schema Fix Bundle — Pick a Page Type, Get Complete JSON-LD Ready for AI Fill
The Schema Validator tells you what's missing. Schema Completeness lets you pick fields one at a time. The Schema Fix Bundle picks a page type and emits the full @graph with every Google rich-result required field pre-populated plus an AI prompt to fill in your content.
How Many Fan-Out Queries Does Your Site Actually Cover?
You generated 40 fan-out queries for your seed keyword. Your site has 80 pages. How many queries have a matching landing page? The scorer fetches every URL in your sitemap, extracts title / H1 / meta, Jaccard-matches against each fan-out query, and reports coverage % plus the orphan-query list you need to publish against.
When Your Entity Graph Splits Three Ways Because of a Naming Drift
Your homepage says 'Acme Corp.' Your blog says 'Acme.' Your About page says 'Acme Corporation LLC.' Your entity graph splits three ways and LLMs ground on whichever version appeared most in training. The checker crawls your site, extracts every JSON-LD Organization / Person / Product and OG tag, and flags the drift.
One Keystroke to Audit Any Page: Installing the jwatte.com Browser Extension
A tiny Chrome / Edge / Brave / Vivaldi extension that routes the current tab to any of the 70+ jwatte.com audit tools with one keystroke. Install in thirty seconds, works offline, no telemetry, source-available.
Isolating IoT Devices On A Consumer Router — VLAN + Per-SSID VPN Segmentation For Beginners
A grey-market streaming stick, a robot vacuum mapping your floors, a doorbell camera on the same Wi-Fi as your laptop — every IoT device is a data leak. This walks through a three-SSID segmentation pattern you can build on any modern consumer router with VLAN + VPN-client support, with examples for NordVPN / Proton VPN / Mullvad and a beginner-friendly visual workflow. The exit countries are yours to pick; the pattern is the same.
Counting AI-Answer Mentions — The Simple Math Behind Share-of-Voice
Share-of-voice in an AI answer is a word-counting problem, not an AI-ML problem. Paste the response, count your brand hits vs competitor hits, compute the fraction. The worksheet does it per response, keeps 25 runs in localStorage for trending, and exports CSV.
Source Diversity — The AI-Answer Metric Nobody Talks About
AI-answer engines cite sources. The extractor pulls every URL from a pasted response, dedupes by domain, classifies each source (own, competitor, Wikipedia, community, news, government), and scores citation diversity. A low-diversity answer is one you could dominate — a high-diversity answer is one where one more citation slot is hard to win.
Broken Link Fix Generator — Turn an Audit Into a Deployable Redirect Bundle
The audit tells you what's broken. This generator emits the deployable fix: stack-specific redirect config, sed patch to rewrite in-HTML references, and an IndexNow payload so Bing picks up the change within minutes.
Auditing Your ads.txt Before a Programmatic Buyer Asks About It
ads.txt and app-ads.txt are the IAB Tech Lab's way of letting publishers publicly declare which ad exchanges are authorized to sell their inventory. Done wrong, DSPs drop you from auctions. Here's how to audit both files, what each line means, and how the jwatte.com ads.txt Audit tool spots the common mistakes.
You've Got The Claude Code CLI Installed — Now What? Plan Mode First, Then CLAUDE.md, Context Hygiene, And The Slash Commands That Matter
Follow-on to the AI Terminal Kickstart. The CLI is the runway; the workflow is the plane. Plan mode first, CLAUDE.md as a living contract, the slash commands grouped by intent (/btw /fork /rewind /compact /loop /schedule /simplify /batch /powerup /insights /debug /claude-api), Skills vs Rules vs Memory, Auto Mode boundaries, and a Day 0 to Day 7 ramp plan.
Your ISP Already Knows When You're Home — How To Take The Surveillance Back And Run Monitoring Yourself
Comcast WiFi Motion turns your router into a presence sensor. The other major ISPs are running similar plays. Here's what each one collects, how to disable it, and how to run uptime/alerts yourself without paying $30/mo to surveillance-adjacent SaaS.
When Google Ignores Your Sitemap Signals
If every URL in your sitemap has priority=0.5, changefreq=weekly, and lastmod=today, Google ignores all three. Boilerplate sitemap signals are noise; meaningful ones are signal — there's a measurable difference in crawl efficiency.
Why WebAuthn / Passkey Posture Audit Exists
Probes a domain for WebAuthn / passkey signaling on the homepage and common endpoints (/.well-known/webauthn, /api/webauthn/options, common auth-vendor paths...
When Your Access Logs Whisper Before They Scream
By the time GSC tells you Googlebot lost interest, you've already lost 3-6 weeks of crawl. Raw access logs see the regression on day one — the trick is knowing what pattern to look for.
Why Web App Manifest Audit Exists
Fetches manifest.json and scores required + recommended fields for PWA installability: icons, name, start_url, display, theme_color.
Why WCAG A11y Audit Exists
WAVE-style accessibility audit — WCAG 2.1 / 2.2 AA + 3.0 draft checks, CSV/PDF/HTML export, AI fix prompt.
Why Voice Cleanup Exists
De-slop AI text — normalize em-dashes, strip slop phrases, break long sentences.
Why View Transitions API Audit Exists
Scans for @view-transition CSS, startViewTransition(), view-transition-name, ::view-transition pseudos, reduced-motion guards.
Byte-Level Plagiarism Detection Misses The Overlap That Hurts Rankings
Copyscape catches exact-match duplication. Google's systems catch paraphrased near-duplicates. The gap is the 5-gram shingle zone — where content looks different to a human but identical to the ranker.
Why URL Structure Hygiene Exists
Internal-link scan: path depth, URL length, capitalization, trailing-slash consistency, session IDs, file extensions, underscores.
Why Trusted Types Audit Exists
Checks require-trusted-types-for CSP + scans for innerHTML/eval DOM sinks. Trusted Types block DOM XSS.
Why Third-Party Script Cost Audit Exists
Enumerates every external script on a page, estimates transfer size + main-thread cost, ranks by impact. Identifies the third party most responsible for CWV ...
Why Subresource Integrity (SRI) Audit Exists
Checks SRI integrity hash coverage on every cross-origin <script> and <link rel=stylesheet>. Missing SRI = supply-chain risk.
Why Speculation Rules API Audit Exists
Parses <script type=speculationrules> blocks. Validates prerender/prefetch rules, eagerness levels, source/where predicates.
When 200 Is The New 404
A page returning HTTP 200 with 50 words of 'no results found' is a soft-404 to Google. It wastes crawl budget, drags site quality scores, and never ranks. The audit catches them in batches and recommends remediation per page.
Why Slug Rename Helper Exists
Scan site for every link to old slug — emits sed patch + 301 redirects.
Why Site-Wide Crawl Sampler Exists
Fetches sitemap, samples 1-25 URLs with template diversity, scores each, shows distribution + worst / best URLs.
Why Sitemap Segmentation Generator Exists
Fetches a sitemap.xml, classifies URLs by type (home / blog / news / product / category / tag / user / video / tool), emits a sitemap-index plus per-type seg...
Why Service Worker Audit Exists
Fetches sw.js, inspects install / activate / fetch handlers, skipWaiting, navigation preload, caching strategies, offline fallback.
The Coverage Gap Between Your Schema And Your Actual Pages
Your LocalBusiness schema declares 22 service areas. You have landing pages for 9 of them. Google and AI grounding systems treat the gap as a credibility problem — you claimed coverage you can't back up.
Check The Update Calendar Before You Blame Your Own Work
When traffic drops, the instinctive response is to audit what you changed. The right first move is to check whether Google rolled an algorithm update. This tool correlates your GSC data against the confirmed update calendar.
Prioritize Title-Tag Rewrites By Traffic Impact, Not Impression Volume
Most SEO rewrite lists are sorted by impressions. That's the wrong signal. The right signal is (impressions × CTR gap vs expected). Small differences in priority ordering create 2-3x differences in monthly traffic.
Why SERP Cohort Audit Exists
Runs 3-10 target queries on DDG, maps which competitors share your ranking cohort, and tells you if you are even in the cohort.
Why SEO Roadmap Generator Exists
Paste findings list; auto-builds Impact × Effort matrix + 30/60/90-day execution roadmap weighted by business type.
Why Ranking For The Wrong Intent Is The Most Expensive SEO Mistake
A product page ranking for a 'how to' query is lighting money on fire. Intent mismatch explains most of the 'my traffic is up but conversions are down' stories. This classifier makes the mismatch visible.
Why Schema Completeness Exists
Generate complete Schema.org JSON-LD graph per page type with linked @id.
Why robots.txt Simulator Exists
Paste a robots.txt, a list of URLs, and a bot. Simulates per-URL allow/disallow decisions per RFC 9309 group precedence rules. Validates before deploy.
When The Same robots.txt Blocks One LLM And Allows Another
GPTBot honors most-specific blocks. ClaudeBot honors the LAST matching block. PerplexityBot ignores wildcards in some patterns. The same robots.txt yields different access maps per bot — and most operators never check.
Why Rich Results Eligibility Audit Exists
Per-type required + recommended field check across FAQ, Article, Product, Recipe, HowTo, Review, Event, JobPosting, Course, LocalBusiness.
Why Review Velocity Drift Precedes Local-Pack Loss
Point-in-time review audits miss the signal that actually matters: the drift. A business going from 4.7 to 4.5 stars over 90 days loses local-pack visibility about 60 days after drift begins. This tool catches the drift early.
Not Every Review Deserves Equal Response Time
An operator with 40 new reviews can't triage by gut feel — some matter urgently, some can wait a week, some don't need a response at all. This tool scores each review on impact × urgency × tone sensitivity and sorts the queue.
Six Ways To Tell A Retriever Your Content Is Current
For time-sensitive queries, retrievers (Gemini Grounding, Perplexity, Claude with web tools) explicitly prioritize fresh content. 'Fresh' isn't inferred — it's signaled. Six specific signals you should emit, every article.
Why Responsive Image Audit Exists
Compares natural image dimensions against displayed width. Flags images shipped 2-8× larger than rendered — the top mobile LCP killer.
Why Response Size Budget Audit Exists
Fetches a page, sums HTML body + CSS + JS transfer, checks against common performance budgets (200KB HTML, 500KB CSS, 500KB JS, 2MB total, Google 15MB crawl ...
Why Reporting API Audit Exists
Checks Report-To, Reporting-Endpoints v2, NEL, CSP report-to wiring, COEP Report-Only coverage.
Why Redundant Alt Text Audit Exists
Finds images where alt duplicates adjacent heading/caption text. Fixes "read twice" screen-reader friction.
Why RAG-Readiness Is The Next Table-Stakes Audit
Every SMB wanting to stand up an internal AI assistant will hit the same wall: their own website can't be ingested cleanly. RAG pipelines demand semantic HTML, passage boundaries, and freshness signals most pages never think about.
Why 'Near Me' Queries Deserve Their Own Rank-Tracking Workflow
Near-me queries behave differently from branded or city-name queries. Proximity dominates. Measuring rank at 0mi / 1mi / 3mi / 5mi / 10mi distance bands reveals the decay curve most SMBs never see.
The New Attack Surface: Competitors Editing What LLMs Say About You
Hostile actors embed hidden instructions on webpages to manipulate LLM summaries. Most payloads are invisible to humans — CSS cloaking, HTML comments, zero-width characters, base64 blobs. This audit scans for the patterns.
Why Prerender vs Hydration Parity Exists
Inspects initial HTML for SPA / CSR patterns: empty SPA root, missing H1, JS-only main content, missing nav links. Critical for Googlebot first-crawl.
Why User Preference Media Queries Audit Exists
Audits prefers-color-scheme, prefers-reduced-motion, prefers-contrast, forced-colors support + color-scheme meta.
Why Preconnect + DNS-Prefetch Hygiene Exists
Audits preconnect, dns-prefetch, preload, prefetch hints. Flags duplicate hints, crossorigin mismatches, and render-critical origins missing preconnect.
Why PQC Migration Plan Generator Exists
Pick the cryptographic surfaces you operate (TLS, SSH, code signing, S/MIME, IPsec, Git signing, validators, wallets, DKIM, DNSSEC, WebAuthn, mTLS, KMS, DB T...
Why Non-Composited Animation Audit Exists
Scans CSS for transitions / animations on non-GPU properties (color, box-shadow, top/left). Finds the repaint-heavy ones.
Why noindex / X-Robots-Tag Conflict Audit Exists
Probes a URL, compares HTML meta-robots directive against HTTP X-Robots-Tag header. Flags silent-kill conflicts where the two layers disagree.
Why Monoclone Gen. Exists
Scaffold an industry site from template + inputs — hero/services/FAQ/schema/sitemap.
Why 'Optimize For AI' Is Too Generic To Be Useful
Gemini pulls 40-60 word self-contained passages. ChatGPT favors definitive claims with numbers. Claude tolerates nuance. Perplexity wants freshness signals. 'AI-optimize your content' is not a strategy — per-model optimization is.
When The New Model Release De-Cites You
Each LLM release re-shuffles citation preferences. A page Claude 3 cited 12 times might be cited zero times by Claude 4 Opus six months later. The model changed; the page didn't. Tracking drift over time tells you when to react and how.
Why Mixed Content Audit Exists
Scans for http:// resources on https:// pages. Detects both active (script, iframe, CSS) and passive (img, video) mixed content.
Why Mega SEO Analyzer Exists
Orchestrator. Runs parallel probes across 7 dimensions (ranking authority, hygiene, AEO / AI retrieval, performance, security, trust, compliance) and returns...
All Local SEO Probes In One Orchestrator
Local SEO is 8+ scattered tools. The Mega GEO Analyzer orchestrates 10 dimensions in one pass — equivalent to BrightLocal Core ($39/mo) + Local Falcon ($24/mo) bundled, free.
All AEO Probes In One Orchestrator
AEO has fragmented into 15+ individual audits. Nobody runs them all in sequence. The Mega AEO Analyzer orchestrates 10 dimensions in one pass and emits a single Health Score with ranked fixes — equivalent to Profound's $499/mo dashboard, free.
Why MCP Server Audit Exists
Probes for Model Context Protocol server publication at /.well-known/mcp-server.json, /.well-known/agent-card.json, /.well-known/ai-plugin.json. Audits manif...
Why Your Site Should Advertise Callable Actions To AI Agents — And How To Start
Agent-native commerce is no longer theoretical. ChatGPT agents, Claude desktop, Gemini copilots all look for /.well-known manifests before they'll call your site. This tool audits what you're advertising and generates starter manifests.
Why Your Local Rank Is A Lie If You Only Measure From One Place
Google's local pack is hyper-local. Your rank at your address is not your rank from a customer's neighborhood two miles east. This post covers why you need a grid and how to build one without paying for Local Falcon.
Why llms.txt Quality Scorer Exists
Fetches /llms.txt, /.well-known/llms.txt, /llms-full.txt. Scores structure, link count, description depth. Flags boilerplate vs useful canonical content index.
Why llms-full.txt Matters More Than llms.txt For Actual LLM Retrieval
llms.txt is a short pointer file. llms-full.txt is the long-form content map LLMs actually consume. Sites that publish both are ingested more reliably and cited more accurately than sites that publish only the short form — or neither.
The Common Crawl Test: Are You Even In The Dataset That Trained The LLM?
Common Crawl is the single biggest upstream input to every major LLM's pretraining. If your domain isn't in it, you're invisible to offline model knowledge. This audit probes CC directly and maps inclusion across C4, FineWeb, RedPajama, OSCAR.
Why The Same Page Costs 30% More To Cite In One LLM Than Another
Different LLMs tokenize the same English content at radically different rates. A retrieval-augmented generation system that pulls your page pays per-token; if your prose tokenizes inefficiently, you compete worse on cost-per-citation.
The Token-Efficiency Audit Every Content Team Ignores Until The Bill Arrives
If you're building a RAG pipeline or letting an in-house AI assistant read your content, token counts are a real cost. Different tokenizers count the same text differently. This tool estimates per-model tokens + dollars per page.
When An LLM Quotes 200 Words Of You Verbatim, Is That Citation Or Republication?
If an answer engine reproduces 150+ words of your content verbatim, the user has no reason to click through. The line between cite-and-link and cite-and-replace is shorter than most publishers realize, and there's a measurable threshold.
Why Live Citation Surface Probe Exists
Probes 18 reference corpora (Wikipedia, Wikidata, G2, Trustpilot, Crossref, OpenAlex, etc.) to show where you are cited and where you are not.
Why Link-Graph Depth Audit Exists
Extracts all outbound links from a page, classifies each domain by authority tier (government / academic / major publisher / known / unknown), scores outboun...
Why Legal Pages Audit Exists
Probes a site for Privacy Policy, Terms of Service, Accessibility Statement, Cookie Policy, Disclaimer, Refund Policy, DMCA agent, and DPA. Checks presence, ...
Why Knowledge Graph + Wikidata Entity Audit Exists
Checks a domain for Knowledge-Graph triggers. Queries Wikidata for existing entity matches, audits Organization schema + sameAs array coverage, recommends st...
Why Keyword Inspection Exists
Inspect the Google top 10 to design your information architecture first.
Why A $0 Invoice Generator Is One Of The Highest-Leverage SMB Tools
The math on invoicing-SaaS is bad for most small businesses. A ledger of $15-19/month add-ons across six tools eats 3% of a $10k/month operator's revenue. A browser-side invoice generator does the 90% case without the subscription.
Why Intrusive Interstitial Audit Exists
Detects patterns Google penalizes on mobile: sticky overlays, full-screen popups, cookie walls that block content, forced newsletter signups.
Orphans Are The Silent Indexation Killers On Most SMB Sites
Orphan pages — pages with zero inbound internal links — are the most common reason SMB content stays in 'Discovered – currently not indexed' purgatory. The fix is internal links from topically-relevant sources, not content rewrites.
Why Internal Link Auditor Exists
Crawl a site; finds every 404 internal link, 301 chain, stale slug + sed auto-rewrite.
Why Inline Event Handler Audit Exists
Scans DOM for onclick/onerror/on* handlers, javascript: URLs, inline scripts, document.write, eval. Blocks strict CSP.
When 80% Of Your Content Is Invisible To Google
Infinite-scroll lists keep users engaged but typically hide 80% of their content from Googlebot. The crawler renders one viewport, sees a scroll trigger, and stops. Most SMB sites with infinite scroll never realize they're indexing only the first 12 items.
Why IndexNow Submission Audit Exists
Checks IndexNow key file at root. Validates filename matches content. Bing/Yandex/Seznam/Naver submission readiness.
Why Image Sitemap Extension Audit Exists
Fetches sitemaps. Verifies xmlns:image namespace, counts image:image tags per URL, scores caption/title/license coverage.
Image SEO Is A Five-Dimensional Audit, Not A One-Line Check
Most Image SEO audits check alt text and stop. Google Images ranks on alt + filename + dimensions + surrounding text + image-sitemap inclusion. A complete audit looks at all five — and 80% of SMB sites only do one.
Why Image LCP Candidate Audit Exists
Identifies the likely LCP image, flags loading="lazy" conflict, missing width/height, no srcset, legacy formats.
Why Identical Link Text Audit Exists
Groups links by visible text; flags groups where same text points to different URLs. WCAG 2.4.4.
The Hyperlocal Terms Your Site Doesn't Mention Are The Queries You're Losing
Every neighborhood name, landmark, school, and ZIP your site doesn't mention is a local query it can't rank for. The coverage matrix is usually bleaker than operators think — and the fix list is cheaper than they fear.
Hreflang Generator — Paste Your Locale Map, Get the Correct Block for Every Format
Most hreflang generators emit one format. This one emits all four — HTML link tags, sitemap.xml xhtml:link blocks, HTTP Link headers for PDFs, and Next.js/Astro framework arrays — with auto x-default and ISO validation built in.
Why HCU Pattern Detector Exists
Scans a URL for Google Helpful Content Update demotion signals: thin content, missing first-party experience, affiliate-without-insight density, AI-boilerpla...
Grounding Is A Different Surface Than AI Overviews — Optimize Separately
Gemini Grounding and Vertex AI Grounding pull from Google Search as source. They weight publisher trust, Wikipedia-linked entities, and Article schema more heavily than AI Overviews. Same site, same content, different optimization.
Why The Un-Answered Question On Your GBP Is A Paid-Ad Tax
Google Business Profile's public Q&A is visible forever. Every un-answered question drives customers to a competitor who did answer. Every user-answered question propagates misinformation. Here's how to catch and fix both.
Why Quarterly Photo Uploads Are Losing To Weekly Ones
GBP photo count matters; photo freshness matters more. Businesses adding 3+ photos per month outperform those with large archives that haven't seen a new upload in six months. The signal Google reads is 'actively operating.'
Why GA4 / GTM Configuration Audit Exists
Detects Google Analytics 4 + Google Tag Manager on a page. Checks Consent Mode v2 integration, IP anonymization, GTM preview-auth token leak, duplicate measu...
Why Form Conversion Audit Exists
Counts form fields, flags unlabeled inputs, missing autocomplete hints, and friction patterns that cut mobile conversion.
Why Font Loading Strategy Audit Exists
Audits @font-face font-display, preload hints, woff2 coverage, Google Fonts vs self-hosted, FOUT/FOIT pattern risks, variable fonts.
Why Featured-Snippet Extractability Audit Exists
Scans a page for Position-Zero-ready structure: question-shaped H2s + 40-60 word answer paragraphs, 3-10 item lists, properly-headed tables, definition patte...
Getting To A Featured Snippet vs Owning One — The Displacement Playbook
Most snippet-optimization advice ends at 'write a 40-60 word paragraph.' That gets you close. Displacing the incumbent takes specific knowledge of their weaknesses. This tool reverse-engineers the incumbent and outputs the replacement shape.
Why E-E-A-T Generator Exists
Person/Org/sameAs/Wikidata/ORCID/rel=me JSON-LD from a single author profile.
Why E-E-A-T Audit Exists
Scores the four E-E-A-T pillars from Person/Org schema + sameAs depth.
Why Edge Cache Effectiveness Probe Exists
Probes a list of URLs, analyzes Cache-Status / CF-Cache-Status / X-Cache / Age headers, reports per-host edge-cache hit ratio, identifies URLs bleeding origi...
Why Early Hints (103) Audit Exists
Detects RFC 8297 HTTP 103 Early Hints + Link: preload/preconnect headers. Typical 100-300ms LCP win.
Why A Byline Without An Entity Behind It Is Worth Zero To Google
Google's E-E-A-T system doesn't evaluate authorship by reading bylines. It evaluates authorship by resolving bylines to entities — Person schema, sameAs links, credentials, external profiles. A plain-text byline is invisible to the system.
When Two Of Your Own Pages Cannibalize Each Other
External duplicates get the headlines; internal duplicates kill more sites. Two of your own pages competing for the same query is a silent ranking killer that most SMBs never measure.
Why DNSSEC PQC Posture Audit Exists
DoH-based audit of a domain DNSSEC chain. Reports algorithm IDs (RSASHA256/ECDSAP256/ED25519 — all classical), DS hash algos (flags SHA-1), NSEC vs NSEC3, KS...
Why CSP Strictness Audit Exists
Deep CSP audit: nonce/hash coverage, strict-dynamic, unsafe-inline, report-to, Trusted Types, wildcard quality.
Why CSP Allowlist Completeness Exists
Cross-references every loaded script/style/iframe against CSP directives. Finds the gap before users see console errors.
Why Crypto Wallet Quantum Risk Audit Exists
Paste a Bitcoin or Ethereum address. Classifies script type (P2PK/P2PKH/P2MS/P2SH/P2WPKH/P2WSH/P2TR or EOA/contract), checks reuse + spend history via public...
Why CrUX Field Data Probe Exists
Real-user Chrome UX Report p75 LCP / INP / CLS / FCP / TTFB via PageSpeed Insights API. The metrics Google actually ranks on.
Why Cross-Origin Isolation Audit Exists
Verifies COOP + COEP + CORP headers for SharedArrayBuffer, high-res timers, WASM threads. Required for advanced web APIs.
Why Critical CSS Inlining Audit Exists
Counts render-blocking stylesheets, inlined <style> bytes, deferred CSS via media=print+onload pattern, font-preload hints.
Why CORS Headers Audit Exists
Deep CORS audit: ACAO * + credentials, Vary: Origin, Allow-Methods, Expose-Headers, Max-Age, Timing-Allow-Origin.
Why Cookie + Storage Drift Audit Exists
Flags pre-consent trackers, document.cookie writes, and localStorage writes in initial HTML — the typical GDPR / ePrivacy failure modes.
Why Content Gap TF-IDF Analyzer Exists
Top-10 SERP TF-IDF analysis. Finds which phrases top rankers use that you do not. Clearscope / SurferSEO core algorithm, in browser, free.
Why Content Decay Audit Exists
Paste URLs + last-modified dates (or GSC Performance CSV). Ranks by staleness, flags content past 12 / 24 / 36 month thresholds, queues refresh priority.
Why Compression Codec Audit Exists
Verifies Brotli vs gzip vs zstd on HTML + subresources, checks HTTP/3 Alt-Svc advertisement, Vary: Accept-Encoding.
Reading Competitor Strategy From Their Google Business Profile Deltas
A competitor adding Saturday hours tells you they're chasing weekend work. A new service category tells you they're expanding. Most SMBs never notice because nobody watches. This tool turns periodic snapshots into a strategic intelligence stream.
Why Competitor Gap Matrix Exists
4-way side-by-side SEO score comparison. Your URL + 3 competitors across 6 dimensions, with gap identification.
Seven New Commercial-Conversion Tools — Checkout, Pricing, Revenue Attribution, About Page, Form Friction, Upsell Surface, Scroll-Depth CTAs
Seventy-five audits on the site optimize for ranking. These seven optimize for revenue. Checkout Abandonment Autopsy, Pricing Psychology Audit, Revenue-Per-URL Attribution Map, About-Page Conversion Audit, Form-Field Friction Estimator, Upsell + Cross-Sell Surface Audit, Scroll Depth + CTA Visibility Map.
Why Code Signing Trust Audit Exists
Probes a GitHub repo, npm package, or website for code-signing trust roots: Sigstore Cosign, SLSA provenance level, in-toto attestations, npm --provenance fl...
Why CMP Compliance Audit Exists
Scans a page for consent-management-platform (Cookiebot, OneTrust, CookieYes, Osano, Termly, iubenda, Didomi, TrustArc). Checks IAB TCF v2.2, CCPA opt-out li...
Why Client Hints Audit Exists
Checks Accept-CH + Critical-CH response headers. Validates UA Client Hints coverage as User-Agent string deprecates.
Why CLAUDE.md Generator Exists
Give Claude Code a permanent project memory — 10-question form, 8 stack presets, outputs drop-in Markdown.
Why TLS Cert Lifecycle Audit Exists
Probes a hostname for cert lifetime (90 / 100 / 397 day buckets), CT-log inclusion via crt.sh, public-key reuse across renewals (the harvest-now-decrypt-late...
Word-Level TF-IDF Gets You Close. Structural Fit Gets You There.
Content briefs from Clearscope and Surfer tell you which words to use. They don't tell you the page SHAPE that wins. Heading depth, schema types, media ratio, internal-link density — aggregate those across the top 10 and you have the canonical template.
Why Cannibalization Audit Exists
Runs a search for a target query + site: operator and identifies when multiple own-site URLs compete for the same query. Recommends consolidation, canonicali...
Why Most SMBs Can't Name Their Break-Even — And What Changes When They Can
Four numbers decide whether a small business is a business or an expensive hobby: fixed cost, variable cost per unit, price per unit, and volume. Most operators can't recite three of them. This calculator makes it a five-minute conversation.
Why Brand-SERP Audit Exists
Probes the top 10 organic results for a brand query. Classifies each (own-site, directory, social, review, negative, knowledge-graph), scores brand-SERP heal...
Why Bing + Copilot + IndexNow Audit Exists
14 signals for Bing Places + Bing Search + Microsoft Copilot + ChatGPT Search citation. Crawlability, CDN-WAF bingbot allow, sitemap.xml, IndexNow key, ai.tx...
Why Batch Compare Exists
Compare up to 10 sites across SEO, GEO, AEO, schema, E-E-A-T, a11y.
When To Disavow, When To Ignore — A Heuristic Playbook
Google's official guidance in 2024-2026 is that most spammy backlinks are auto-discounted. But some still hurt. Knowing which is which — without paying $100+/month for Semrush or Ahrefs — is a scoring problem the tool solves free.
Why Author Authority per Article Exists
Scores an article on 8 authorship signals: byline, Person schema, author link, photo, bio, rel=author, datePublished, dateModified.
Why API Secret Leakage Audit Exists
Scans HTML + JS for 20+ secret patterns: AWS, Stripe live, Twilio, OpenAI, Anthropic, GitHub PAT, Slack tokens, JWTs, private keys.
Why Site Analyzer Exists
Launch + Agency + Network + Hosting checks in a single report.
Why ai.txt Generator Exists
Spawning-style AI training opt-in/out covering every major AI crawler in the shared bot registry — custom paths, robots.txt companion.
Pretraining-Visible vs Retrieval-Only Is A Real Distinction
Every LLM has a training cutoff. Pages published before it are baked into the model's default knowledge. Pages published after are only reachable via live retrieval. Your content strategy should know which bucket each page sits in.
The First-Mover Window On AI-Answer Queries Is Still Open — Here's How To Find It
For queries where an AI Overview already fires, displacement is expensive. For queries where no AI answer has fired yet, you can BE the canonical answer. Finding the second bucket and investing content there is the single highest-ROI AEO play of 2026.
AI Referral Traffic Without a Paid Analytics Subscription
Paid AI-traffic tools bill $99-$299/month to count how many visitors arrive from ChatGPT, Perplexity, Gemini, or Claude. The numbers live in your access log already. The new parser buckets them by AI referrer, classifies human-follow-through vs bot-fetch UA, and runs entirely in your browser — paste never leaves the page.
Why AI Model Recommender Exists
Describe a task + constraints → ranked AI model recommendation across Anthropic, OpenAI, Google, Meta, Alibaba, DeepSeek, Mistral (text + code), plus image (...
Why AI Model Fit Audit Exists
Paste the prompt + the model you used → get a fit assessment (category match, capability gaps, context overflow, cost/latency mismatch) plus better-fit candi...
When LLMs Get Your Brand Wrong: How To Measure And Correct Hallucinations
Every major LLM makes factual errors about real businesses. This post covers how the drift happens, why it persists, and how a free browser-side tool can catch it before a customer does.
Why AI Crawler Log Analyzer Exists
Paste server access logs. Classifies hits by AI crawler (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, Applebot-Extended, Bytespider, etc.), reports per...
When Robots.txt Allows GPTBot and Cloudflare Still Returns 403
Three layers decide whether an AI bot reaches your content: robots.txt / ai.txt, meta robots, and CDN bot-protection. The first two are easy to audit. The third silently 403s bots even when the first two say allow. The auditor scores all three across every major AI crawler.
Why AI Content Disclosure Audit Exists
Checks for visible AI-disclosure, schema.org creativeWorkStatus, SoftwareApplication author, C2PA signing references, AI-use policy page.
Why Cited-At-All Is The Wrong AEO Metric — Position Is The Whole Game
Being cited in an LLM answer feels like a win, but being cited fifth barely moves traffic. First-position citation in an AI answer drives 3-5x the click-through of mid-list positions. Tracking position longitudinally is the actual game.
The Description LLMs Use For You Is Your First AI-Mediated Impression
Before a customer clicks any search result, an LLM may have already described your brand to them. 'Established but outdated' vs 'innovative and modern' is a positioning difference that happens before your site loads.
Why Unattributed Facts Are The Next Thing LLMs Filter Out
A page claiming '7% of SMBs fail in year one' with no citation is indistinguishable from a page making the number up. LLMs are increasingly filtering against unattributed numeric claims — attribution coverage is an emerging ranking signal.
When AI Traffic Hits A Wall Instead Of A Funnel
An AI engine cites your blog post. A user clicks through. The page is a 1500-word article with no email-capture, no CTA, no clear next step. They leave. Citation traffic that doesn't convert is the most-overlooked AEO problem of 2026.
Per-Model Accuracy Tracking Without A $500/Month Subscription
The value of scheduled LLM monitoring isn't the monthly bill. It's the cadence. This tool makes the cadence free: paste + score + log + trend across ChatGPT, Claude, Gemini, Perplexity over weeks and months.
Why Your Site Needs A Sitemap For Actions, Not Just Pages
Traditional sitemap.xml tells crawlers what to READ. Agents need to know what to CALL. A parallel machine-readable list of your site's invocable actions — book, quote, verify, order — is the next discoverability layer.
Broken Actions Silently Destroy Your Agent Visibility
Declaring callable actions to agents is worthless if the endpoints are broken, slow, or auth-mismatched. Agents that hit failing endpoints deprioritize your site — quietly — for every future invocation. This probe catches failures before they cost you.
Why 404 Page Quality Audit Exists
Fetches a guaranteed-missing URL and scores the 404 page: HTTP status, useful copy, home link, search bar, branded layout, canonical hygiene.
Mega GEO + Mega AEO — Fix / Audit / Learn Pills On Every Finding
The two Mega orchestrators that were stuck at 'here's a list of problems' now behave like Mega Analyzer and Mega Security Analyzer — every finding carries Fix / Audit this / Learn pill buttons that deep-link into the specialist tool pre-filled with your URL.
The Four Questions I Wish I'd Asked Before Picking an AI Project
I spent a couple of years picking AI projects the way a developer would — what's technically interesting? That was the wrong question. The right one has four parts, and the non-coders shipping products in 2026 are the ones who answered them first. Here's the short version, with worked examples from courts, hospitals, and road crews.
Claude Design — What Anthropic Labs Actually Shipped, Who It's For, And The One Feature That Makes It Different From Figma
Anthropic Labs launched Claude Design on April 17 2026. It builds a design system from your codebase, iterates on prototypes by chat and sliders, and hands finished work to Claude Code. Here is what it does, three real use cases, where it breaks, and when adopting it is worth the bill.
HTML Watermark Tokens for Clone Detection. A Simple Pattern Every Site Should Use
How to embed a stable, Google-searchable watermark token in every page of your site so clones announce themselves within 24 hours of being indexed.
What Your Car Knows About You, and What Automakers Sell It For
A 2024 congressional investigation documented Hyundai selling data from 1.7 million vehicles to insurance broker Verisk for about $1.04 million. McKinsey projects connected-vehicle data services at $450 to $750 billion globally by 2030. Here is what is actually going on, who the brokers are, and what the Electronic Frontier Foundation recommends you do about it.
Video Schema + Transcript Audit — why your YouTube embed doesn't rank
YouTube owns video search. But your page can still win the text-on-a-page-with-video layer — if it has VideoObject schema, a crawlable transcript, and captions. Most pages with embedded video have none of the three.
Third-Party Data Leakage Audit — 22 trackers classified by PII risk
GDPR, CCPA, and CPRA all require disclosure of third-party data transfers. Most sites have 6-12 trackers loading and no idea what data they send. This tool classifies 22 known trackers by PII risk tier and flags the ones that cross the compliance line.
Soft-404 Detector — the pages that return 200 but signal 'not found'
A soft 404 is a URL that returns HTTP 200 but the body says 'page not found', 'no results', or is thin-to-empty. Google demotes and eventually deindexes them. They're invisible to naive crawlers. This tool crawls your sitemap and surfaces them.
Auditing Serverless Functions for the 25 Posture Failures That Cost You Money or Get You Owned
Serverless platforms shift cost + security to the request shape. Origin allowlists, rate limits, CORS-with-credentials, SSRF, secret-in-log, fetch timeouts. 25 checks that prevent quota burn, account takeover, and stack-trace leaks across Netlify / Cloudflare Workers / Vercel / Lambda / Fly / Deno / Render / Railway / DigitalOcean.
Which SERP features do your queries trigger? The matrix tells you in 60 seconds
SEMrush and Ahrefs charge $100+/mo to tell you which SERP features (Featured Snippet, PAA, Local Pack, Shopping, Video Carousel) your target queries trigger. This fires the queries at DDG + Bing, detects 12 feature types, and shows you which ones are the biggest opportunities.
Your longitudinal SEO dashboard was already in localStorage — I just drew it
Every Mega SEO Analyzer run writes the score to your browsers localStorage. The SEO Trend Dashboard reads it and renders per-URL line charts with improving/regressing/stable classification. No server, no account, no export/import needed.
Turning the audit tool into a monitoring service — scheduled, notifications, no server
ContentKing charges $195/month to watch your site and ping you on regressions. The Scheduled Auto-Monitor does it in the browser with setInterval + localStorage + Notifications API. Keep a tab open, get alerted when your score drops 5 points or more.
The roadmap needed a board — drag SEO findings from To-do to Done
Linear charges $8/seat. Trello caps free boards. This is a 4-column SEO-specific kanban that lives in your browsers localStorage. Bulk-import audit findings as cards, drag across columns (To do / Doing / Blocked / Done), export JSON, watch your SEO backlog clear.
Schema Test-Case Generator — 16 valid JSON-LD templates, required fields enforced
Most broken schema is broken because someone copy-pasted an example missing required fields. This tool emits 16 valid JSON-LD templates — Article, Product, FAQPage, HowTo, LocalBusiness, Recipe, Event, Organization, BreadcrumbList, VideoObject, Review, JobPosting, Course, SoftwareApplication, Book — with every required property filled in.
Making JSON-LD visible — the schema graph visualizer I wish Schema App didn't charge for
Every page has a structured-data graph hidden inside its JSON-LD blocks. Schema App charges $29/mo to draw it. This tool parses your JSON-LD, flattens it by @id references, renders the full node/edge graph as interactive SVG. Finds orphans, missing @context, weak connectivity.
Real User Monitoring without a server — RUM that stores to localStorage
Sentry Performance charges $29/mo + per-transaction fees. This generates a self-contained RUM snippet that captures LCP / INP / CLS / FCP / TTFB on every pageview and writes to your sites localStorage. No endpoint, no third-party data collection, no recurring cost. Paired with an import-JSON dashboard.
The SPA audit gap, solved by paste — no serverless Chromium required
Single-page apps inject content after hydration. My static-HTML audits miss it. Prerender.io costs money. Headless Chromium in a serverless function costs even more. The Rendered DOM Paste Audit is the pragmatic free path — you paste outerHTML from DevTools, the tool audits the post-hydration DOM, and optionally diffs against the server HTML.
Readability Analyzer — Flesch-Kincaid, ARI, and the passive-voice tax
Google's Helpful Content system reads for readability directly. Sentences averaging 24+ words, passive-voice density above 15%, and Flesch-Kincaid grade level above your audience all correlate with demotion. This tool measures all three.
APE, RACE, CREATE, SPARK — The Four Prompt Frameworks Every Prompt Should Match One Of
By 2026, four named prompting frameworks — APE, RACE, CREATE, SPARK — dominate the literature. Each is a checklist of the elements an LLM needs to produce a good output. The new auditor scores any prompt against all four, tells you which one it's closest to, and emits a rewrite that closes the gaps.
The pillar-cluster pattern shows up in every AEO article — so I built the audit
A 3,000-word pillar linking to eight focused subtopic articles that link back. It's the architecture AI retrievers treat as authoritative. Most blogs aren't structured this way, and it isn't obvious from reading a single page.
Tracking AI-search visibility — are Perplexity, you.com, and Bing AI citing you?
Google rank tracking tells you where you appear on the SERP. AI answer engines cite sources in their answers and those citations are the new SERP. This tool queries Perplexity, you.com, and Bing AI for your target queries and reports whether your domain shows up in their answer sources.
The Legal Pages Generator. Privacy, Cookies, Terms, Accessibility, Disclaimer, Refund, DPA in One Pass
Indie sites need seven or eight legal pages and most ship two. The new Legal Pages Generator produces a baseline set tailored to business type + jurisdictions + third-party services, with a master AI customize prompt.
The link-graph visualizer Sitebulb charges $149/mo for, free in your browser
Crawls your sitemap, extracts internal links per page, runs iterative PageRank, renders an interactive SVG graph. Shows which pages are hubs, which are orphans, and exactly where PageRank flow is starving content. No signup, no crawler bill.
Generic positioning loses to specific positioning in agent-driven search
An AI agent looking for a CRM for an automotive dealership will not pick 'CRM for businesses of all sizes'. It'll pick the page that says, out loud, 'CRM for automotive dealerships with 10-50 seats'. This tool audits whether your page says that out loud.
Pre-deploy SEO feedback in the browser — the HTML Sandbox Playground
JSFiddle and CodePen are great for JS but dont check SEO fundamentals. This sandbox audits your HTML live as you type - title, meta, canonical, OG, Twitter, JSON-LD, alt, word count. Save up to 20 named drafts locally. Test before you deploy.
Hreflang Cluster Graph — visualize reciprocity and catch every broken pair
Hreflang works only if every locale variant references every other locale variant — reciprocally. One missing reciprocal and the entire cluster collapses silently. This tool crawls your hreflang cluster and renders the (in)completeness as a graph.
Hreflang Audit — Fetches Every Page in the Cluster and Builds the Backlink Matrix
Most hreflang auditors only inspect one URL. This one fetches every alternate page in the cluster, cross-checks backlinks, flags malformed ISO codes, and surfaces canonical-vs-hreflang conflicts that silently invalidate Google's cluster recognition.
IBM published a GEO playbook — I turned it into a page audit
IBM published a Generative Engine Optimization playbook in April 2026. Core thesis: from keywords to prompts, from links to citations, from websites to ecosystems, from traffic to answer eligibility. This tool scores your page across ten GEO signals drawn directly from the playbook — Q&A format, answer directness, conversational tone, chunk-ideal paragraphs, schema, DOM simplicity, cite markup, authorship, freshness, structure.
Cross-Domain Entity Consistency — Wikidata, sameAs, and the entity graph
AI retrievers resolve brands to entities in Wikidata. If your site's Organization schema disagrees with your Wikidata record — or you have no Wikidata record — you're a stranger to the entity graph. This tool reconciles both sides.
I stopped generating fix prompts. Meet the Code-Diff Patch Generator.
Every other audit tool outputs a prompt you paste into Claude. This one skips the LLM loop and writes the HTML patch itself — unified-diff format, before/after blocks, ready to paste into your template. How to use it, why I built it, and when it still loses to prompts.
Canonical & Redirect Graph — visualize every chain SEO crawlers hate
Redirect chains cost crawl budget and dilute PageRank. Canonical conflicts tell Google two different stories. This tool crawls your sitemap and renders the tangle as an SVG graph — the shape of the problem is the first useful diagnostic.
Broken Link + Decay Scanner — find every 404 your page points at
Outbound links decay. The average 5-year-old article has 10-25% of its external references dead, redirecting to parked domains, or quietly 302-ed into spam. This tool fetches every outbound link in parallel and reports the damage.
Anchor-Text Manipulation Detector — SpamBrain catches EMA ratios you don't feel
Google's SpamBrain doesn't just count backlinks — it profiles your exact-match anchor distribution. One moderately-trafficked page I audited had 43% EMA for the same commercial phrase. This tool surfaces the ratio before SpamBrain does.
Share-of-Voice in AI Answers Without a $99/Month Subscription
Paid AI-visibility platforms charge $99-$499/month to do one thing: run 25 prompts through each LLM and count your brand mentions. The new Prompt Pack tool generates the prompts and the scoring worksheet for free — you run the prompts by hand in the free tiers, log the results in a CSV, and get the same share-of-voice data without the subscription.
I read 14 AEO articles and found 5 signals my 254 tools didn't measure
After parsing 14 AEO articles from SearchScore, Semrush, Neuralic, The Conductor, and broworks, I found five specific content signals none of my existing audits explicitly checked. So I built one more.
Findings lists don't persuade — so I built an audit narrator
Audit tools output checklists. Executives want a narrative. This tool takes any findings list, categorizes by topic + severity, weights by your business model, and produces an executive summary with prioritized next-step sprints. Deterministic, no LLM tokens.
The agentic web is the new shopping surface — an audit for it
Microsoft launched AI Max + Universal Commerce Protocol + Copilot Checkout in April 2026. Google is pushing agentic shopping through Merchant Center. Perplexity Shop is live. Your product pages need to be discoverable, trusted, and transactable by AI agents — not just ranked on Google. This audit scores your readiness.
Every new security audit tool I built this month, and the CSP that doesn't block Clarity
CSP strictness, Trusted Types, Reporting API, SRI, CORS depth, COOP/COEP, secret leakage, inline event handlers, mixed content, CSP allowlist completeness — what each one catches and why I needed it.
Ten new performance audit tools that exist because Lighthouse wasn't enough
Early Hints (103), CrUX field data, compression codec, critical CSS, font loading, preconnect hygiene, Speculation Rules, LCP candidate, size budget, View Transitions — each one catches a regression Lighthouse reports vaguely.
The compliance audit stack — CMP, GA4/GTM config, cookie drift, AI disclosure, CCPA/GDPR
Six tools that check what a CMP audit actually checks — consent mode, blocking-mode enforcement, pre-consent cookie writes, IAB TCF v2.2 signaling, Global Privacy Control handling, and EU AI Act Article 50 readiness.
Every new AEO tool for 2026 — llms.txt quality, MCP server audit, AI content disclosure, author authority
AI retrievers (Anthropic, Perplexity, Google-Extended) don't cite your site the way Google crawls it. These tools check the surfaces they actually use: llms.txt, MCP servers, author apparatus, citation surface across reference corpora.
The accessibility + UX audit tools I built after Lighthouse stayed polite
Redundant alt detection, identical link text, prefers-color-scheme + reduced-motion + forced-colors, non-composited animations, intrusive interstitial patterns, form conversion friction. What each one flags, and why Lighthouse didn't.
Netlify Application Firewall vs Cloudflare Bot Management, and the Double-CDN Trap
Netlify ships Application Firewall, Log Drains, and Analytics. Cloudflare ships Bot Management, Turnstile, and WAF rules. Putting both in front of your origin is not free. Here's the decision matrix.
Why I turned Mega SEO Analyzer into a v2 and stopped paying $99-$500/mo for Ahrefs-adjacent tools
v2 added real-user CrUX field data, site-wide sample crawl, 4-way competitor gap matrix, localStorage trend baseline, and a 30/60/90 roadmap. Free, no signup, no tracking. Here is why I built each piece.
Lighthouse caught 5 things my other audit tools had missed — so I built new ones
Desktop scored 83 performance, mobile 67, LCP 13.1s on slow 4G. I fixed the site and built 5 new audit tools so the same issues surface in every future audit. Here is each fix and each tool.
DMCA Takedowns and Terms of Use for Developers. A Practical Walkthrough
What a valid DMCA takedown notice actually has to contain, where to send it, what to write in your Terms of Use, and the safe-harbor mechanics that make the whole process work.
The Content and Tool Protection Playbook for Indie Developers
Function guards, watermarks, terms, clone monitoring, log drains, attribution. The complete defense stack we ship on jwatte.com, in the order we'd ship it on a new site.
Content Operating Systems Explained. Why Sanity, Strapi, Contentful, and Payload Are Replacing the Monolithic CMS
A Content OS separates content from presentation: structured schemas, an API-first content lake, and a separate authoring studio. The pattern fixes WordPress's coupling problem, but introduces deploy-time, schema-migration, and CDN concerns of its own. When to use one, when to skip.
The Task You Should Never Have Been Doing: Notes on Handing Work to a Computer-Use Agent
Computer-use agents change the delegation question. It isn't 'can the AI write this code?' anymore — it's 'should I be doing this task at all, given that an agent could run it while I work on something actually mine?' Four quick tests to decide, three workflows where I said yes, three where I said no, and the small set of habits that keeps it all safe.
Picking the Right AI Model in 2026. Why GUI and CLI Are Different Tools, and Where Image Generators Diverge
Pros and cons of every major model in the 2026 lineup. Claude family, GPT-5 / o4 / Codex CLI, Gemini 2.5 + CLI, Llama 4, Qwen 3, DeepSeek V3 + R1, Midjourney v7, Ideogram v3, DALL·E 3, Flux 1.1 Pro, Stable Diffusion 3.5, ElevenLabs, Deepgram. Why GUI prompts and CLI agents are different products. Why we ship a Claude Code kickstart file. When to reach for which image generator.
The Three WordPress Themes Worth Installing in 2026. Kadence, GeneratePress, Astra
Most WordPress performance problems start at the theme. Three lightweight themes keep showing up as the correct default in 2026. Here's how they differ and which one fits which site.
Four WordPress URLs That Should Never Be Publicly Reachable. wp-login, readme, license, install
Four URLs ship with every WordPress install. Three of them expose your exact core version to scanners. One of them can take over your site if left in the wrong state. Here's what to do about each.
Why I Built Site Migration Capture. A Consent Banner I Forgot Cost Me a Client
I migrated a client from WordPress to Eleventy and forgot their consent banner. Two weeks later a GDPR complaint arrived. The client left. This tool exists so nobody else makes the same mistake on the same migration.
Nine Files That Shouldn't Be Publicly Reachable On Any Website
Regardless of whether you run WordPress, Shopify, a static site, or a custom app, nine files commonly ship to production by accident and leak credentials, code history, or internal structure. Here's what each one is and how to block it.
Pricing Pages and Careers Pages. The Two Transparency Signals Most Sites Hide
Hidden pricing tanks B2B conversion, and a missing careers page reads as a dead company. Two of the simplest trust signals a site can ship, and they cost almost nothing to add.
Your Yelp Listing Is Probably Bleeding Star-Rating Context. 11 Signals the Yelp Audit Checks
Yelp still drives discovery for restaurants, services, and local trades in major US metros. If your /biz/ URL is wrong, your public rating is unlinked, or your review count can't be parsed from HTML, you're losing Copilot and Apple Maps citations — not just Yelp search. This audit names the 11 signals that matter.
A WordPress and WooCommerce Audit That Speaks WP, Not Generic SEO
WordPress has its own specific failure modes that generic SEO tooling misses: plugin pileup, abandoned themes, legacy jQuery, stacked render-blocking CSS, WooCommerce cart-fragments on every page. This tool reads WP-native signals and emits a WP-native fix prompt.
A WordPress Hardening Audit That Speaks to the Server, Not Just the Admin Panel
A WordPress-specific security audit. Probes XML-RPC, REST API user enumeration, admin and login exposure, debug.log and wp-config backup reachability, directory indexing, and security-header coverage. WooCommerce PCI notes where applicable.
A 30-Second Uptime Diagnostic Beats a Paid Monitor When You Just Need to Know If The Site Is Up
Paid uptime monitors run every 5 minutes from 10 locations and cost $30-$100 a month. For a quick diagnostic of 'is my site reachable right now and how fast is it,' a single-shot server-side probe is enough.
HVAC, Plumbing, Electrical, and General Contractors. Six Site Signals That Move the Needle
Trades websites live or die on local-pack visibility. Six signals separate a site that converts from one that costs money to host. Here's what they are and how to audit them in one pass.
Before You File The LLC Or Print Business Cards, Spend 20 Minutes On USPTO Trademark Search
A $500 to $1,500 trademark attorney's clearance opinion is the right answer for anything you're serious about. Before you pay the attorney, spend 20 minutes on a pre-screen so the attorney's hours are focused on real risk, not first-pass searching.
FTC Click-to-Cancel Takes Effect in 2026. California and New York Already Enforce. Here's Your Autorenewal Audit
FTC Click-to-Cancel rule + California BPC 17600-17606 + New York GBL 527-a + Virginia, Colorado, Illinois, Oregon, and New Jersey autorenewal statutes all apply to US subscription businesses. Most SMBs are compliant with zero of them. This tool scores the 12 signals that matter across every major statute.
Every State's Business-Entity Search In One Place. Vendor Vetting, Competitor Research, Your Own Good-Standing Check
All 50 state Secretary of State business-entity searches plus DC. Verify LLC / corp status, registered agent, filing history. Use cases: vendor vetting, competitor research, confirming your own good standing.
An A-F Letter Grade For Your SSL Config. Because 'TLS 1.2 with SHA256 AES-GCM' Isn't A Sentence For Most Owners
SSL Labs gives you a 400-line technical report. Most SMB owners want a letter grade. This tool rolls up protocol, cipher, cert expiry, HSTS, OCSP, and PQC readiness into A / B / C / D / F.
TCPA Violations Run $500 to $1,500 Per Text. Carriers Block Unregistered Senders. Here's How to Audit Your SMS Program
US business SMS is now governed by three overlapping layers: TCPA (federal statute, $500-$1,500 per violation), CTIA (carrier industry body), and A2P 10DLC (carrier-level registration). Most SMBs comply with one and fail at the others. This tool checks all three.
Capture Before You Migrate. Why the New Site Should Boot Up Functionally Equivalent
Most migrations start from a blank template and slowly rebuild what the old site had. This tool captures the old site's settings footprint first, then emits a generator-specific spawn prompt so the new site boots equivalent from day one.
A Restaurant Site Audit That Actually Understands Menus, Hours, and Reservations
Restaurant sites break in different places than SaaS sites. Menu schema Google reads. Opening hours that match Google Maps. A reservation path. A delivery path. This audit covers the restaurant-specific layer your general SEO tool misses.
Query Fan-Out — Why AI Search Doesn't Actually Run Your Keyword
Google AI Mode, ChatGPT Search, and Perplexity don't retrieve your keyword. They fan it out into 10-60 sub-queries behind the scenes, blend the answers, and cite whoever showed up in enough buckets. Here's what fan-out looks like — and the new tool that maps it for any seed keyword.
Post-Quantum Cryptography for Small Sites. A Practical Migration Path
Classical RSA and ECDH break when a large quantum computer arrives. That computer is not here yet. The encrypted traffic attackers are recording now will still matter when it is. This tool probes your server for hybrid PQC key exchange and emits a migration plan.
A Platform-Aware Audit. Shopify, Squarespace, Wix, Webflow, Ghost, BigCommerce
Most audits treat every site the same way. A Shopify store breaks in different places than a Squarespace portfolio. This tool detects the platform and runs the checks that actually apply.
PCI SAQ-A: The Lightest Compliance Path For Ecom SMBs, And Why You're Probably Not Eligible For It
SAQ-A is the 22-question PCI self-assessment most small ecom merchants qualify for. One misconfigured analytics tag on the checkout page bumps you to SAQ-A-EP (130+ questions) or SAQ-D (380+). Most owners don't know which they're actually in.
Your Payment Processor Assigned You The Wrong MCC. Here's How To Check
MCC mismatch costs SMBs 0.3 to 1.5 percent on every transaction. On $500K annual sales that's $1,500 to $7,500 a year in fees you shouldn't be paying. Most owners don't know their MCC, let alone whether it's right.
Seven Security Layers In One Scan. Mega Security Analyzer
One URL. Seven security layers scanned in parallel. TLS version and cipher, post-quantum hybrid KEX, HTTP security headers, CSP strictness, DNS email-auth, MITRE ATT&CK tactic mapping, CWE and OWASP pattern scan. 3-probe false-positive suppression.
A Medical Practice Audit That Knows The Difference Between SEO And HIPAA
Medical sites carry a compliance layer the tool can't audit (HIPAA, state board rules). It can audit the public-facing signals that show up in real medical-board advertising reviews: provider credentials, insurance, booking, privacy notice, tracking technology.
Google, Apple, Yelp, Bing. The Four Local-Presence Platforms Every US Small Business Needs to Own
Most SMBs optimize Google Business Profile and stop. That covers about 55% of US local search — the remaining 45% runs through Apple Maps (iPhone default), Yelp (home services, restaurants, medical, legal, auto), and Bing (which also powers Microsoft Copilot, ChatGPT Search, DuckDuckGo, and Yahoo). Here's the four-platform audit most SMBs never run.
Eight Local Directories Actually Move Your Rankings. Here's How To Check Coverage
Citations across Yelp, BBB, Yellowpages, Angi, Foursquare, Manta, Superpages, MapQuest still drive local-pack rankings. Most SMBs are listed on two and missing six. This tool surfaces the gaps.
LLMs.txt Validator — Why Most Published llms.txt Files Silently Fail
Your site has an llms.txt. A retrieval engine just fetched it. Did it parse cleanly, or did the engine silently skip half of it because the H1 was missing, the URLs were relative, or two sections had the same name? The new validator runs all 12 structural checks and tells you exactly what to fix.
A Law Firm Site Audit Built for SEO and Bar-Compliance at the Same Time
Law firm sites carry a compliance layer no other vertical does. State bar rules govern what you can say. This audit covers SEO structure (LegalService schema, practice areas, jurisdictions) and scans for risky language (guaranteed, specialist, best) that triggers ethics review.
Index Coverage Delta — The Diff Between What GSC Sees and What You Actually Publish
GSC reports 'Discovered — currently not indexed' without telling you why. This tool crawls your site, diffs against sitemap.xml, and produces the exact punch list of orphans, ghosts, canonical conflicts, and redirect chains blocking indexing.
Independent Hotels Lose 15-25% of Revenue to OTA Commissions. Here's the 16-Signal Audit That Recaptures It
Booking.com and Expedia take 15-25% of every booking they drive. Independent hotels and motels recapture that by strengthening Google Hotels eligibility, direct-booking incentives, channel-manager coverage, and review aggregation. This tool scores the 16 signals that determine cross-channel revenue.
Your Google Business Profile Is Leaking Revenue. 18 On-Site Signals That Fix It
Forty percent of local leads come from Google Business Profile. Most SMBs under-optimize it because they audit their website and their GBP separately. This tool audits the 18 on-site signals that reinforce GBP ranking and lift review conversion, and shows exactly which dashboard field to match each to.
CAN-SPAM, TCPA, CCPA, GDPR. Four Statutes Governing Your Email and SMS Footers
CAN-SPAM penalties run $51,744 per violating email. TCPA penalties run $500 to $1,500 per violating text. This tool audits your footer for the required elements across four statutes.
Your Password Has Probably Been Leaked. Here's How To Check Without Trusting a Random Tool
Have I Been Pwned hosts over 11 billion leaked credentials. The safe way to check your own password uses k-anonymity: SHA-1 hash locally, send only the first 5 characters, compare the rest in your browser. Never type your real password into a tool that sends it across the internet.
Contrasting Three Competitors Without a Subscription Seat. How the Competitor Contrast Tool Works
Most competitor-intel tools cost 100 to 500 a month and lock results behind a seat. This one fetches each URL, extracts keywords, frameworks, benefit cues, and pricing hints, then hands you a scrape plan and a comparison prompt for Claude or ChatGPT.
ADA Web Accessibility Lawsuits Hit 4,500 Filings in 2024. Most Targets Are SMBs Under $25M
Title III web-accessibility complaints are a multi-year growth industry and the target profile has shifted to small businesses. This tool scores your site on the 12 signals plaintiff firms put at the top of their demand letters, then produces a remediation plan that actually reduces exposure.
Static-Site Generator Deployment Hardening. Eleventy, Hugo, Astro, Gatsby, Next, Jekyll, SvelteKit, Nuxt
Static-site generators claim 'just deploy the build output' but each generator has its own set of deployment gotchas that show up in production. Eleventy deploys the wrong folder. Hugo leaks drafts. Next.js ships source maps. Ten common issues, per-generator fixes.
SpeakableSpecification: The Schema Tag That Tells AI What to Quote
SpeakableSpecification markup tells Google, voice assistants, and AI tools exactly which text on your page is worth quoting. Here's how to implement it and why it matters for AI citations.
Testimonials, Case Studies, Logo Strips, and Awards. Which Ones Actually Move the Needle
Four kinds of social proof turn up on almost every serious B2B page, and most sites pick the weakest three. Here's what testimonials, case studies, logo strips, and awards each actually do, and which ones AI engines cite.
Three Server-Side Probes That Let Browser-Based Audits See Past the Browser
A browser can't query DNS. A browser can't inspect a TLS handshake. A browser can't rerun a request three times to filter CDN-race false positives. Three small server-side probes fix all three gaps.
Your Server Header Is Telling Attackers More Than You Think. Apache, Nginx, Caddy, IIS, Windows Server, Red Hat, Oracle
The Server: and X-Powered-By headers tell attackers your exact version of Apache, nginx, Caddy, IIS, ASP.NET, Express, or Oracle, plus frequently the underlying OS. Then CI/CD config leaks (terraform.tfstate, .gitlab-ci.yml, Dockerfile) finish the job. Here's what each leak reveals and how to block it.
The Six Security Headers Every Site Should Ship in 2026. And The Three That Actually Matter
Security header recommendations change every three years. Here's the 2026 baseline: what to ship, what to skip, what the modern COOP / CORP pair does, and why CSP is the one you can't get away without.
When Live Chat and Booking Widgets Actually Earn Their Screen Space
Conversion widgets pay off on pages with strong buying intent and a short decision path. On info pages they just get in the way. Here's how to tell which page you're looking at.
Cookie Consent Vendors. Which One Actually Matches Your Legal Exposure
Nine consent vendors run about 90 percent of the banners on the internet. Each one sits at a different intersection of price, legal rigor, and site-speed impact. Here's which one fits which site.
What Actually Fixed My Claude Code Sessions (It Wasn't a Prompt Trick)
I spent months chasing prompt templates. None of them mattered. The four changes that turned Claude Code from a frustrating toy into something I ship with are almost boring — a written brief before any code, a memory file in the repo root, live data instead of pasted snapshots, and a spec tight enough that the model can only be wrong one way.
Numbers, Badges, Licenses, and Where You Got the Numbers. The Four Claims Signals That Actually Rank
AI answer engines cite specific numbers, visible compliance badges, license numbers, and sources. Generic marketing language loses every time. Four signals that take an afternoon to add and are worth the engineering time.
Dual-Audience Navigation and Honest AI Messaging. Two Signals Most Sites Skip
Sites serving two distinct audiences usually try to serve both with one navigation and end up serving neither. AI products that hide their human-review process lose enterprise buyers. Two informational signals worth the engineering.
A $10,000 Crypto Portfolio Made 24.8 Percent in 30 Days With An AI Agent. Here's How To Actually Back-Check It
Medium article claims an AI agent turned 10,000 USDC into 12,480 USDC in 30 days on Solana. Here's how to verify the math, what prompts and services were used, and what the real after-cost, after-tax return looks like for a regular reader.
Eight More WordPress Audits That Pay for Themselves. XML-RPC, User Enumeration, Heartbeat, and Five Other Quiet Problems
The Mega Analyzer and the WP+Woo audit catch the obvious WordPress problems. Eight more checks live one layer deeper: XML-RPC exposure, REST API user enumeration, wp-emoji, jQuery Migrate, Heartbeat, block library CSS, WooCommerce payment gateways, and review schema.
YouTube Channel Audit — Trailer, Playlists, Description, Cadence
Channel trailer, playlists, subscriber count, About description, banner. Five signals that decide whether a one-time viewer becomes a subscriber. Audit a channel home in three minutes.
X Is Gated Now — Audit What Signed-Out Visitors Can Still See
X quietly gated most profile data behind login. The public preview still leaks enough to audit — bio, profile image, website link, location, verification. This tool grades what your public card actually shows.
The .well-known Directory Audit — Probe 13 Files, Fix the Gaps
Audit /.well-known/ for the files that matter in 2026 — security.txt, agent-card.json, ai-plugin.json, change-password, webfinger — and generate copy-paste replacements for anything missing.
TikTok Profile Audit — The Signals That Decide If You Get A Link In Bio
Bio hook, link-in-bio (unlocked at 1K followers), niche hashtags, pinned video, avatar, verification. Audit a TikTok profile against the signals that TikTok's own discovery algorithm weighs.
A Syndication Plan in 30 Seconds — Canonical, Rewrite, Stagger
Paste an article URL. Pick the platforms. Get a canonical-tag plan, per-platform headline brief, UTM-tagged cross-post URLs, and a staggered schedule that keeps the original ranking.
Substack Audit — The Seven Signals Before A Reader Hits Subscribe
Subtitle, cover image, subscriber count, paid-tier setup, recent post, podcast option. Audit the public home of any Substack publication against the signals that drive the first-time-visitor subscribe rate.
Speakable Schema, Generated From The Page You've Already Shipped
Paste any URL. Get a SpeakableSpecification JSON-LD block that marks the headings and direct-answer paragraphs AI assistants should read aloud — generated from the HTML you already published.
Single Site Gen — The AI Build Prompt That Already Passes Our Audits
Single Site Gen writes a site-build prompt with every best practice our audit tools already score for — so the first deploy passes the Mega Analyzer without a round of fixes.
Reddit Community Audit — Vet A Subreddit Before You Invest Promotion Time
Subscriber count, active-user count, description depth, karma (for user profiles), and engagement signals. Audit a subreddit before you post into it — or your own profile if you're doing community-led marketing.
Why 'Take A Deep Breath' Makes LLMs Smarter — The Research Behind Prompt Enhancement
Detailed personas, step-by-step framing, stakes language, and self-evaluation — each one has a paper behind it. This post walks through the research so you know why the Prompt Enhancer does what it does.
Why Most Patreon Pages Convert At Under 1%
Tier pricing, welcome video, recent-post cadence, perk clarity, patron count visibility. The six signals that decide whether a Patreon visitor becomes a paying member — or bounces in three seconds.
MemPalace, Explained — And A Generator For Your First Setup
MemPalace is a local-first, zero-cloud AI memory system that stores raw conversation verbatim in ChromaDB, organized in a spatial hierarchy. This post explains how it works and ships a setup generator.
Twelve Local Directories, Thirty Minutes, One Coverage Check
Yelp, BBB, Bing Places, Yellow Pages, Foursquare, Nextdoor, Apple Maps, Facebook, Angi, HomeAdvisor, Thumbtack, TripAdvisor. Twelve directories that shape your local citation graph. Find out which ones you're missing in three minutes.
Audit Your LinkedIn Newsletter Like It's A Growth Funnel (Because It Is)
Most LinkedIn newsletters are an abandoned Substack draft with a bad cover image. This audit scores cadence, word-count sweet spot, cover, hook density, hashtags, and CTA — with a rewrite prompt.
Kubernetes Readiness — The Twenty Signals That Separate A Prod Cluster From A CV
20-signal Kubernetes readiness checklist: resource requests & limits, security context, probes, network policies, secrets, RBAC, backups. Plus an AI remediation prompt that produces patch YAML for whatever you're missing.
ItemList & Carousel Schema Without Writing It By Hand
Paste a list of URLs. Get valid ItemList or Carousel JSON-LD with correct position fields for Google rich results — in Summary, Article, Product, or Course flavors.
Your Instagram Bio Is A 150-Character Landing Page. Audit It Like One.
Bio hook, link-in-bio, category badge, contact button, profile photo, posts count. Nine signals decide whether Instagram is doing anything for your business. This audit surfaces which ones are missing.
Your Google Maps Listing Decides The Phone Call. Audit It.
Rating ≥ 4.2, ≥ 20 reviews, phone visible, address correct, recent photos, accurate hours. Eight signals decide whether a Google Maps searcher calls you — or the business next door.
GitHub Repo Visualizer — The Shape Of Your Code, At A Glance
Paste any GitHub repo. See stars, forks, open issues, default branch, license, language composition, top 10 contributors, and recent releases on one page.
Your GitHub Profile Is An E-E-A-T Asset. Score It Like One.
AI models weigh GitHub profiles as sameAs-grade identity signals. This audit scores a user, org, or repo against the 14 signals that drive topical authority — README depth, topic tags, license, releases, contributors.
GBP Competitor Research Without A $300/Month Tool
17 signals, 5 competitors, one worksheet. Do the GBP competitive research that local-SEO SaaS sells for $300/month — in 30 minutes with a structured checklist and an AI brief that translates gaps into action.
The Only AEO KPI You Actually Own — GA4 LLM Referral Segment
Pick the AI answer engines to track and get a ready-to-paste GA4 audience, Looker Studio filter, GTM trigger, and BigQuery SQL. The AEO KPI that lives in your analytics, not a paid dashboard.
Bulk-Tag a Thousand Keywords by Funnel Stage — The Funnel Keyword Audit
Paste a keyword list, get every row tagged by funnel stage and intent, plus a distribution chart, a CSV export, and a content-brief AI prompt that turns the list into a page-by-page plan.
Facebook Pages Still Matter For Local Businesses — Audit Yours
Profile photo, cover image, hours, phone, website link, rating, about section. Facebook Pages are quietly one of the largest review platforms online. This audit scores the signals that drive local discovery.
Docker For A Small Business — Generate The Compose File You Actually Need
Pick the services your small-business stack runs. Get a best-practice Dockerfile (multi-stage, non-root, health-checked) and a docker-compose.yml with volumes, networks, Traefik auto-HTTPS, and the env plumbing you'd forget.
Google Discover Readiness — The 14 Signals That Actually Drive Placement
Discover is Google's quietly enormous discovery surface. Most articles miss one of three things — hero image width, max-image-preview:large, or NewsArticle schema. This audit catches all three in one pass.
Who Signed Your Image? A Content Credentials (C2PA) Checker
Scan a page or single image URL for C2PA / Content Credentials manifests. Find out which of your images carry a provenance chain and which don't.
Docker vs docker-compose vs Kubernetes — A Small-Business Decision Tree
Interactive stage-by-stage diagrams of the three containerization tiers. Plain Docker, docker-compose, Kubernetes. See what each layer does for you — and which tier your business actually needs.
Cameo Profile Audit — Price, Sample, Response, Rating, Reviews
Five signals decide whether a Cameo visitor books. Sample video, price transparency, response-time promise, star rating, review count. Audit yours in under a minute.
Your BookBub Profile Is A Conversion Funnel. Fix It In 10 Minutes.
Twelve signals decide whether a BookBub visitor becomes a follower. Most author profiles fail on three of them. This audit surfaces exactly which three.
Ship A Bluesky Custom Feed On A Cloudflare Worker
Compose a Bluesky / AT Protocol custom feed from keyword include/exclude rules, account allow/deny lists, and language filters. Outputs a deployable Cloudflare Worker.
What 'Hosting On IPFS' Actually Means — A Step-By-Step Visualizer
Interactive diagrams showing exactly how ENS, Handshake, and Unstoppable Domains resolve to content, and how IPFS and Arweave hosting actually store your data — step by step, phase by phase, with the commands at each step.
Blockchain Stack For A Small Business — When It Pays, When It Doesn't
Compose a blockchain-native stack — decentralized DNS (ENS, Handshake, Unstoppable), decentralized hosting (IPFS, Arweave), crypto payment rails (BTCPay, Coinbase Commerce, Stripe Crypto, Lightning) — for a small business. Honest trade-offs, no hype.
Apple Business Connect — The GBP Clone Everyone Forgets
Apple Maps covers roughly 30% of US users by default. Apple Business Connect is the GBP equivalent — with Showcases, Action items, Indoor Maps, and native iOS handoffs. Most businesses don't claim it.
App Store Listing Audit — The Seven ASO Signals That Decide Installs
Icon, description hook, rating, rating count, last update, screenshots, keyword coverage. Audit your App Store or Google Play listing against the ASO signals that gate install conversion.
Repo Strategy — Public, Private, Branch Approvals, And The Privacy Tier Most People Skip
When to make a repo public. When to keep it private. How branch approvals actually work. The alternatives to GitHub (GitLab, Gitea, Codeberg, Sourcehut, Forgejo). And the legitimate privacy-focused hosting that lives on Tor / Proton — with the legal context most writers skip.
Passage Retrieval Is The New SEO — Why Your Page Score Doesn't Match Your AI-Citation Rate
AI answer engines chunk your page into ~150-token passages and cite the individual chunks, not the page. A page with a great overall SEO score still loses citations if every passage is too thin, pronoun-dependent, or missing named entities. The new Retrieval tab scores every paragraph on the jwatte.com tools.
Mark It N/A: Dismiss Audit Checks That Don't Apply to Your Site
Every SEO audit surfaces checks that don't apply to your site — rel=author on a company site, Wikidata for a single-location shop, ORCID for a non-academic author. The jwatte.com tools now let you mark those Not Applicable, persist the decision, and exclude them from the AI fix prompt.
llms.txt Structural Validation — Beyond Presence, Does Yours Actually Match the Spec?
Most llms.txt files on the web are just a flat dump of URLs. The llmstxt.org spec calls for an H1 title, a blockquote description, H2 sections, and structured link lists — the exact shape AI retrieval engines expect. Here's the canonical structure and the validator that enforces it.
Hreflang Must Be Bidirectional — Why Google Demotes Multilingual Sites With One-Way Links
If your English page declares a Spanish alternate, the Spanish page must declare English back. One-way hreflang silently demotes both pages in Google. The fix is a reverse link; the audit is a single pass through every declared alternate.
Export Your Audit: Make Every SEO Scan Reproducible
The Mega Analyzer and Site Analyzer now let you export a full audit as JSON, import it later, and regenerate the AI fix prompt — without re-fetching the site. One-click export means scans become shareable, diff-able, and archivable.
AI Terminal Kickstart — Set Up Claude Code, Codex, ChatGPT CLI & GitHub Copilot On A Fresh Box
Step-by-step install for Claude Code, OpenAI Codex, ChatGPT CLI, GitHub Copilot CLI and Netlify CLI on macOS (Homebrew), Windows (PowerShell) and Linux (Bash). One-shot scripts plus copy-paste commands for every dependency, AI framework, and auth step.
Does Your robots.txt Agree With Your ai.txt? The Silent AI-Crawler Misconfig Most Sites Have
Four different files tell AI crawlers whether they can read your site. When they disagree, each bot resolves the conflict differently — GPTBot may honor robots.txt, Perplexity may honor ai.txt, Google-Extended may honor X-Robots-Tag. The audit you need takes 30 seconds and catches the misalignment every SMB site has.
The AI Posture Audit — From Per-Bot Matrix to One Master Prompt
The AI Posture Audit evolved from a single-purpose crawl-directive checker into a full discovery-surface audit. It now fetches 13 identity + policy files, validates the llms.txt structure, and emits one master prompt covering every finding — so you paste once into Claude or ChatGPT and get a consolidated fix plan, not six separate prompts to juggle.
Top AI CLIs — How To Feed Them The Prompts Our Generators Build
Claude Code, Gemini CLI, aichat, OpenAI CLI, llm, and continue. Seven command-line AI tools and a concrete pattern for feeding each one the mega-prompts our site builders and analyzers generate.
Why I Built the WCAG Audit Tool (and Why You Probably Need It More Than You Think)
Over 5,000 ADA website lawsuits were filed in 2025. A Portland yoga studio paid $11,500 for something that takes two hours to fix. I built a free WCAG 2.1/2.2 AA scanner so small business owners can check themselves in fifteen minutes. Here's what it does, why it exists, and the three-step test you can run right now with no signup.
Why I Built a WCAG 2.1 / 2.2 AA + 3.0 Draft Accessibility Audit Tool (and How to Fix Every Issue It Finds)
A free in-browser accessibility audit that mirrors WAVE WebAIM — covers WCAG 2.1 AA baseline, the six new 2.2 AA outcomes, the 3.0 draft readability outcome, exports CSV / HTML / PDF, and hands you an AI fix prompt. Here is why it exists, what it finds, and how to fix every issue yourself.
Newsletter Swap Matchmaker: Fair-Value Calculator, Partner Criteria, and Three Outreach Templates
A new free tool that calculates a fair-swap value for your newsletter (opens, CPM, CPC), surfaces eight partner-fit criteria, and generates cold, warm, and paid outreach templates plus a one-page swap brief.
Claim Your Coined Framework: The Framework Origination Signal Generator
A new free generator that produces DefinedTerm + CreativeWork JSON-LD, a Wikipedia stub, an Internet Archive submission checklist, a canonical outline, and a press-release template so you can be credited as the origin of a framework or method you coined.
Entity Citation Radar: See Every High-Authority Source That Should Cite You
A new free tool that maps sixteen high-authority sources (Wikipedia, Wikidata, Internet Archive, Goodreads, OpenLibrary, Google Scholar, Zenodo, ORCID, Crunchbase, and more) against your book, author, brand, or framework — with a visual radar and a submission checklist for every gap.
Introducing the AI Citation Readiness Audit: Score Your Content for Perplexity, ChatGPT, and Claude
A new free tool that scores any article for how likely Perplexity, ChatGPT, Claude, Gemini, and Google AI Overviews are to cite it — across 14 on-page signals.
The Tools Nav Trap: When Your Dropdown Promises a Page That Doesn't Exist
A 'Tools' dropdown in the nav that never ships a /tools/ landing page is a broken signal. Users expect it. Crawlers expect it. Netlify's catch-all sends it to 404. Here's how I keep catching this and how to fix it in ten minutes.
Score Ten Sites Across Every Dimension in One Run — The Mega Batch
The Mega Batch runs the Mega Analyzer's six-bucket convergence across up to 10 URLs at once. Score matrix, weakest-link tag per row, and a mega batch AI fix prompt that covers the whole portfolio.
One URL, Every Audit in One Pass — The Mega Analyzer
The Mega Analyzer converges every jwatte.com SEO tool into a single site audit. SEO, schema, E-E-A-T, voice, mobile parity, performance, AI-search — six bucket scores, one grade, one copy-paste AI fix prompt.
sitemap-image.xml: The Second Sitemap Most Sites Skip and Multimodal AI Notices
A regular sitemap tells crawlers which pages exist. sitemap-image.xml tells them which pictures live on which page. Here's how to structure it and when shipping it actually matters.
Sitemap Delta: Finding the URLs Your Site Links To That Are Not in Your Sitemap (and Vice Versa)
Crawl your site, diff what you link against what sitemap.xml lists. Three buckets fall out. Here's what each one means and how to fix them.
HSTS, CSP, and the Security Header Bundle: The Trust Signal You Send With Every Response
Six response headers, one deploy, measurable ranking and AI-citation improvements. Here's the bundle I ship on every site and the Netlify config to do it.
Netlify force = true and the Redirect Loop You Ship By Accident
Adding force = true to a Netlify redirect rule feels harmless. On slash-normalizing rules it creates a self-referential redirect loop that curl -L hits the 50-redirect limit on. Here's what's actually happening and when force is safe to use.
max-snippet, max-image-preview, max-video-preview: The Robots Meta Line Most Sites Forget
Google caps snippet length at 160 characters by default. AI answer engines respect the same cap. One meta tag removes it. Here's the line and where to put it.
llms.txt, llms-full.txt, and the .well-known Mirror: What They Are and Why Every Site Should Ship All Three
llms.txt is how you hand AI retrieval engines a clean, curated map of your site. Here's what goes in it, how llms-full.txt differs, and why you should also serve /.well-known/llms.txt.
JSON Feed Alongside RSS: Shipping /feed.json as Well as /feed.xml
RSS is the lingua franca of syndication. JSON Feed is the modern sibling that newer aggregators prefer. Here's the spec, the differences, and why I ship both from the same source.
humans.txt: The One-File Transparency Signal That Still Matters in 2026
humans.txt was invented for the web in 2011 as a credits file. In 2026 it quietly doubles as an E-E-A-T and authorship-verification signal that AI citation engines parse. Here's why and how to ship it.
geo.position, LocalBusiness Schema, and og:locality: The Location Signal Triplet
Apple Maps, Bing, voice assistants, and location-aware AI all read the same three location signals. Miss any one and the local-business page is partly invisible. Here's the triplet.
Footer Completeness: Why Every Sitemap URL Needs At Least One Footer Link
A page in your sitemap that no internal nav links to is an orphan. Search engines deprioritize orphaned sitemap entries because authority has no path to flow to them. Here's the footer rule I follow on every site I run.
dateModified, Updated Stamps, and article:modified_time: The Three-Layer Freshness Signal
AI citation engines weight fresh content. Three freshness signals in three different places is how you prove content is current without gaming it. Here's the layering I use.
Cross-Site Link Hygiene: Why Relative Links Are the Enemy of a Multi-Site Network
When you run six book sites, twenty satellite sites, and one tools hub under one author, relative markdown links become the number one source of 404s. Here's the real case that cost me three deploys to trace and the registry pattern I use now.
<cite> and <blockquote cite=>: The Semantic HTML Most Sites Forgot
Two HTML elements from the 1999 spec that AI citation engines read as source-quality signals. Most sites render quotes with a div and a class. Here's why that is a missed signal.
Every Tag Link Is a 404 Until You Build the Tag Pages
Eleventy blog templates love to emit /blog/tag/SLUG/ anchors on every post. If you never set up tag pagination, every one of those links is a 404. Here's the detection pattern and the template you need to fix it in ten minutes.
How Your Copy Reads — Measuring Voice Against the Pages That Rank
The free Voice & Tone Analyzer at jwatte.com measures reading grade, sentence shape, passive voice, and CTA density for any page — and lets you compare against the Google top 10. Find out why the ranking pages read differently than yours.
Who Owns the AI Overview, Featured Snippet, and PAA for Your Query — A Free Tracker
The free SERP Feature Coverage Tracker at jwatte.com detects every feature firing on a Google SERP — AI Overview, featured snippet, People Also Ask, video carousel, image pack, knowledge panel, sitelinks — and surfaces which hostname owns each one.
Mobile-First Indexing Is Indifferent to What Looks Good on Desktop — Here's How to Check
The Mobile vs Desktop Parity Check at jwatte.com fetches any URL with both a desktop Chrome and an Android Chrome User-Agent, then diffs the HTML, headings, schema, and link count. Catch hidden mobile-only content gaps that Google sees but you don't.
Find Orphans and Hubs With a Free Internal Link Graph
The free Internal Link Graph at jwatte.com crawls your site from a seed URL, charts the topology, and surfaces orphan pages (no inbound internal links) plus hubs (high-authority concentration points) — no account, no paid crawler, runs in the browser.
Harvest Every FAQ From the Google Top 10, Deduped and Ready to Implement
The free FAQ Harvester at jwatte.com pulls FAQPage JSON-LD + accordion questions + question-shaped headings from every top-ranking page for your target query. Deduplicates, clusters by theme, and outputs a ready-to-paste FAQPage schema block.
Track When Competitor Pages Change, What They Add, and Who's Catching Up
The free Content Velocity Monitor at jwatte.com snapshots any set of URLs today. Re-run later and see word-count deltas, new and removed H2s, schema additions, and publish-date changes. Browser-local storage — no monitoring service needed.
Blog Syndication Playbook: How to Cross-Post to X, Medium, and LinkedIn Without Losing SEO
A step-by-step playbook for syndicating one blog post to X, Medium, and LinkedIn in under 15 minutes per post — with the canonical-URL settings that protect your search traffic and a $24/month automation stack that does most of it for you.
Running Keyword Inspection Across a Network of Sites Without Losing a Day
When you own a network of sites, gap analysis at the per-site level doesn't scale. Here is how to use the free Keyword Inspection tool at jwatte.com to run coordinated keyword research across a portfolio of properties. Pairs with The $100 Network methodology.
Launching a New Site? Run Keyword Inspection Before You Write the Homepage
The free Keyword Inspection tool at jwatte.com lets you see what the Google top 10 ranks for before your site even exists. Use it to build your information architecture, not to patch it later. Pairs with The $97 Launch methodology.
The Fastest Keyword Gap Analysis an Agency Can Run in 2026
A walkthrough of the free Keyword Inspection tool at jwatte.com for agency teams running client SEO: paste a target term, the Google top 10, and your client's URL, and deliver a keyword, heading, and schema gap report in under two minutes. Pairs with The $20 Agency methodology.
Why Every Website Needs an AI Agent Card in 2026
AI agents are crawling the web autonomously. An agent-card.json file in your .well-known directory tells them who you are, what you offer, and how to interact with your site. Here's how to deploy one.
Best Personal Finance Books in 2026 (And Why Most Lists Are Wrong)
Every 'best finance books' list recommends the same titles from 1997. Here's why those lists fail in 2026 — and which books actually address inflation, AI, energy mandates, and the insurance crisis.
The Trap Series: Which Book to Read First (And Why It Matters)
Six books. 2,611 pages. But you don't need to read them in order. Your starting point depends on where you are right now. Here's a decision tree to find your Book 1.
If You Liked Rich Dad Poor Dad, Read These Next
Rich Dad Poor Dad changed how you think about money. These 6 books give you the specific strategies, tax math, and exit plans Kiyosaki never covered.
13 Personal Finance Books Worth Reading in 2026 (Most Lists Get This Wrong)
The 13 best personal finance and FIRE books in 2026 — from wealth building to tax strategy to real estate to starting a business. Includes The W-2 Trap, Rich Dad Poor Dad, and more.
I Escaped the Rat Race. Here's the Exact 6-Book Framework I Used.
Ready to escape the rat race? Here's the complete framework — understand the trap (W-2 Trap), build an exit ($97 Launch), promote it ($20 Agency), avoid real estate mistakes (Condo Trap + Resale Trap).
You Have $0 and No Connections. Here's How to Build Wealth Anyway.
No trust fund? No inheritance? Here's how to build wealth from nothing — using tax strategy, low-cost business models, and smart real estate decisions. A 5-phase framework.
You Have $97 and a Laptop. Here Are the Books That Turn That Into a Business.
Want to start a business but have no capital? These 8 books show you how — from $97 startups to bootstrapping to the tax advantages of business ownership.
Why W-2 Workers Stay Broke (Even Making Six Figures)
Making six figures but still broke? The problem isn't your budget — it's structural. Here's why W-2 workers lose wealth to currency devaluation and taxes, and what to do about it.
I Wrote 380 Pages on Condo Economics. The Answer Is Almost Always No.
After writing a 380-page book on condo economics, here's the honest answer: for most buyers in most markets, condos are a bad investment in 2026. Here's why.
Should You Build or Buy a House in 2026? The 25-Year Math
Is it cheaper to build or buy a house? After analyzing all 50 states, the data shows a $400K resale costs $318K-$506K more than a $400K new build over 25 years.
Stop Reading Finance Books in Random Order. Here's the Sequence That Actually Works.
Overwhelmed by FIRE books and financial independence reading lists? Here's the optimal reading order for 2026 — from understanding the problem to building solutions to protecting your wealth.
