Generate an ai.txt file that declares how AI training crawlers may use content from your domain. This is the training opt-in/out signal (analogous to robots.txt for crawling) popularized by Spawning and adopted by OpenAI, Cohere, Anthropic, and others as a courtesy. Complement it with llms.txt (LLM context) and robots.txt (search crawling).
Emerging training / search-crawler user-agents. Tick whether each may train on your content. Boxes start checked where the default stance would allow that bot.
ai.txt is a proposed standard (Spawning, 2023) not yet formalized by an RFC. Compliance is voluntary. Pair it with Cloudflare's "AI Scrapers and Crawlers" rules and with robots.txt User-agent blocks for defense-in-depth. Spawning's spec →
Serve at https://yourdomain.com/ai.txt (root level, same as robots.txt). Content-Type: text/plain; charset=utf-8. Some crawlers also honor /.well-known/ai.txt — serving both is the safest play.