Every major LLM has a training cutoff. Pages published after the cutoff aren't in pretraining — they're visible only via retrieval / browsing. Pages published before are "baked into" the model's knowledge. This audit maps your publish dates against current model cutoffs.
Read the story behind this tool: Pretraining-visible vs retrieval-only is a real distinction →
One per line: URL | publish_date (YYYY-MM-DD). Date of first publication (or first significant publication) per URL.