llms-full.txt is a single Markdown file at a site's root with the full text of its docs, so AI tools and coding agents can load everything in one request.
What is llms-full.txt?
llms-full.txt is a plain-text Markdown file, usually served at /llms-full.txt, that contains the complete content of a website’s documentation in one document. It lets an AI tool or coding agent load an entire docs set in a single request instead of crawling page by page. It complements llms.txt, which lists pages as links with short descriptions.
The llms.txt proposal, published by Jeremy Howard of Answer.AI in September 2024, specifies only the llms.txt file. The “full” variant came from implementers. Mintlify announced in November 2024 that it would automatically host both /llms.txt and /llms-full.txt for documentation sites on its platform, and many developer-tool companies followed. So llms-full.txt is a widely copied convention, not a formal standard.
How it works
A typical implementation:
- Generated from source. Docs platforms build the file automatically from the same Markdown that renders the site. Mintlify’s docs say each page appears with its title, source URL, description and full Markdown content.
- Single fetch. A developer pastes the URL into an assistant, or an agent fetches it, and the whole docs set lands in the model’s context.
- Paired with an index. llms.txt gives a map for selective fetching; llms-full.txt gives everything at once.
| File | Contains | Best for |
|---|---|---|
| llms.txt | Links and one-line descriptions | Navigation, selective fetching |
| llms-full.txt | Full page text in Markdown | Loading complete docs into context |
From SEO to AEO and GEO
The nearest SEO predecessor is the XML sitemap: a file that tells machines what exists on your site. A sitemap only lists URLs for crawlers to fetch and index. llms-full.txt hands over the content itself, already stripped of navigation, scripts and layout, in a format a language model reads easily.
What changed is the consumer. Search crawlers index pages and rank them. Coding assistants and agents want the answer in context now, often inside an IDE. What carries over is the discipline: one canonical source, kept current, with accurate facts.
The impact on AI search visibility is unsettled. Google says no special AI text files are needed for AI Overviews or AI Mode. No major AI search engine has documented using llms-full.txt as a citation or ranking input. Its proven use is developer experience: making your docs easy to drop into a model.
What to do about it
- Prioritize if you sell to developers. DevTools, API and security products with public docs benefit most.
- Generate it, do not hand-write it. Build it from your docs source so it never drifts from the live pages.
- Exclude what should not be there: internal notes, deprecated versions, customer-specific pages.
- Keep it reasonably sized. If it exceeds what common models can hold, split by product or section.
- Link it from llms.txt and mention both on your developer portal.
- Do not expect traffic from it. Measure success by whether assistants answer questions about your product correctly.
Common mistakes
- Treating it as a ranking hack. There is no documented ranking effect.
- Letting it go stale. An outdated file teaches agents wrong API usage.
- Dumping marketing copy into it. Agents need precise, factual documentation.
Related terms
- llms.txt
- Context Window
- AI Crawler
- AI Agent
- Model Context Protocol (MCP)
- Answer engine optimization guide
Sources
- The /llms.txt file proposal, llmstxt.org
- llms.txt and llms-full.txt, Mintlify documentation
- Simplifying docs for AI with /llms.txt, Mintlify blog
- AI features and your website, Google Search Central
Frequently asked questions
Is llms-full.txt part of the llms.txt standard?
No. The llms.txt proposal published at llmstxt.org by Jeremy Howard in September 2024 specifies only the llms.txt file. llms-full.txt is a convention that documentation platforms, notably Mintlify, began generating in late 2024. Many developer docs sites now publish both, but there is no formal spec for llms-full.txt.
Does Google use llms-full.txt?
Google has not said it does. Its AI features documentation states there are no additional requirements to appear in AI Overviews or AI Mode and that you do not need new machine-readable files or AI text files. Treat llms-full.txt as a convenience for tools and agents, not a search ranking factor.
How large should llms-full.txt be?
There is no limit in any spec, but very large files may not fit in a model's context window or may be truncated by tools that fetch them. If your docs are huge, keep llms-full.txt focused on current, public pages, or offer section-level files alongside it.
