An llms.txt manifest tells ChatGPT, Perplexity, Claude, and Gemini which pages on your site to ground citations against. Sites with a valid /llms.txt show a 0.55 correlation with AI-citation probability (Princeton 4-factor GEO framework). Build yours in 90 seconds.
Canonical answer pages — your most authoritative, evergreen content.
Worked examples, case studies, demos — concrete things models can cite.
Supporting context that helps but is not required for grounding.
Spec Readiness
0/100
Preview (first 8 lines)
# Your Site
What
A plain-text Markdown file at /llms.txt that lists your most authoritative pages, organized into Docs, Examples, and Optional.
Why
Retrieval-augmented LLMs (Perplexity, SearchGPT, Gemini) face a context-window budget. llms.txt tells them which URLs to prioritize when grounding answers about your site.
How
Pick 6-12 pages that are self-contained and answer real questions. Skip your homepage, about page, and category indexes — they don't carry standalone meaning.