What is llms-full.txt?
llms-full.txt is an optional companion to llms.txt: one markdown file holding the full text of a site's key pages, so a model or agent can read them in a single fetch. llms.txt is a short reading list. llms-full.txt is the reading itself: every key page's full text in one markdown file. It suits documentation that coding agents load whole, and there is little evidence that answer engines use it to decide who to name.
Where does llms-full.txt come from?
From documentation hosting rather than the llms.txt proposal. The proposal itself, by Jeremy Howard at llmstxt.org (September 3, 2024, modified August 10, 2026), defines /llms.txt and suggests markdown versions of pages; it does not define llms-full.txt.
Mintlify, a documentation host, says it originally developed llms-full.txt in a collaboration with Anthropic, which wanted a cleaner way to feed its whole documentation to language models (Mintlify, May 9, 2025). Mintlify's docs describe the file as combining an entire documentation site into a single file, each page with its title, source URL, description and full markdown (Mintlify docs, read Oct 15, 2026).
How is llms-full.txt different from llms.txt?
One is an index, the other is the content:
| File | Holds | Size | Best for |
|---|---|---|---|
| llms.txt | A summary and links to key pages | Short | An agent deciding what to read |
| llms-full.txt | The full text of those pages | Can be very large | Loading a whole doc set in one fetch |
Do AI engines cite llms-full.txt?
Almost never. In our citation samples from August 23 to October 5, 2026, the 129 sites among the 300 most cited that publish an llms.txt were cited 6,343 times; one of those citations pointed at the file, an llms-full.txt cited by Perplexity on October 2 (the post).
Google states that you do not need machine-readable files, AI text files, markup or Markdown to appear in Google Search (Google, updated July 10, 2026). Treat the file as a convenience for agents, not a visibility lever.
What decides citations instead, in a real example?
The pages engines read for the question. In the public Tesla example (Oct 4, 2026), Google AI Overviews cited youtube.com 20 times and reddit.com 15 times across 25 buyer questions, and tesla.com once.
No file on a brand's own site changes which videos and threads an engine reaches for. Earning a place on the pages already cited, and keeping your own pages readable, does.
Should you publish an llms-full.txt?
Publish one if agents read your docs; otherwise it is optional:
- Yes for developer docs and APIs, where coding agents load whole doc sets.
- Keep it generated from the same source as your pages, so it never goes stale.
- Link it from your llms.txt so agents can find it.
- Check it serves as plain text, not an HTML page at that path.
Common questions about llms-full.txt
Is llms-full.txt part of the llms.txt standard?
llms.txt is itself a proposal, not a standard, and the proposal at llmstxt.org does not define llms-full.txt. The full-text file is a convention documentation hosts adopted.
Does llms-full.txt improve AI visibility?
There is no evidence that it does. One of 6,343 citations in our sample pointed at one, and Google says no AI text files are needed for its AI features.
How big can llms-full.txt be?
As big as your docs, which can exceed what a model takes in one go. Agents that need less can follow llms.txt links instead.
Does Proofsource publish one?
Yes. Our llms.txt ends with a pointer to llms-full.txt, the full text of every page, for tools that want everything in one fetch.
Related terms
- llms.txt
llms.txt is a proposed markdown file at the root of a website that tells language models what the site is and lists the pages most worth reading.
- robots.txt
robots.txt is a plain-text file at the root of a website that tells crawlers which paths they may fetch, under the Robots Exclusion Protocol (RFC 9309).
- Schema markup (structured data)
Schema markup, or structured data, is code on a page, usually JSON-LD using the schema.org vocabulary, that states facts about the page and its subject in a form machines can read.
- Answer engine optimization (AEO)
Answer engine optimization (AEO) is the work of getting your brand named, and your pages cited, in the answers AI assistants give.
Worked example: Tesla, public AI answers, Oct 4, 2026: Proofsource's first scan of Tesla, 25 buyer questions and 5 brand questions, one answer per question on each of four engines. A small sample from one brand, shown as an example, not a benchmark.
See where AI leaves you out, and what to fix.
See which buyer questions name your brand, who gets named instead, and what to fix. 200 answers. 25 Prompts/Question. Top AI engines. 2 days. Free. No card.
Google AI OverviewsGoogle AI Mode
