In one sentence
LLMs.txt is a short Markdown file at /llms.txt that hand-picks the pages you want AI crawlers to read first.
Why it exists
Language models ingest a lot of text but reason over narrow context windows. An LLMs.txt is a curated table of contents: "if you want to understand this site, start here." It reduces the work the crawler has to do, and it raises the chance that the pages you care about land inside the model's retrieved context.
The standard is proposed by the AI and web community. It is not ratified, not standardized through a formal body, and adoption across major AI crawlers is mixed. Support is growing. Some engines honor it, others ignore it today. The downside of publishing one is essentially zero, which is why many sites ship it anyway.
Minimal format
# Site Name > One-line description of the project or brand. ## Main section - [Title](https://example.com/url): Short description. ## Secondary - [Title](https://example.com/other): Short description. ## Optional - [Title](https://example.com/maybe)
For a pre-built version, use the free LLMs.txt generator. It produces a well-formed file in under a minute.
How to track its impact
Ship the file, then watch your citation tracking for a shift in which URLs AI engines cite for your head buyer prompts. Give it weeks, not days. Crawlers re-index on their own cadence, and AI engines weight freshness differently.
Frequently asked
- Is LLMs.txt an official standard?
- No. It is a proposed, community-driven convention. Adoption across major AI crawlers is mixed.
- Does it actually improve AI visibility?
- Evidence is inconsistent. The downside of shipping one is essentially zero. Measure with citation tracking.
- Where do I put it?
- At
/llms.txton your site root. Same location pattern asrobots.txt.