llms-full.txt
llms-full.txt is a companion convention to llms.txt: the complete documentation of a site concatenated into a single markdown file served at /llms-full.txt, so an agent can load the entire docs surface in one request. Where llms.txt is an index that routes agents to individual pages, llms-full.txt is the full text itself, ready to place directly into a model's context window.
The one-request property is the point. An agent that needs deep coverage of an API — every endpoint, every parameter, every error — can fetch a single URL instead of walking a link graph, and documentation platforms increasingly generate the file automatically alongside llms.txt. The Model Context Protocol's own documentation site serves one, which is a reasonable signal of where docs tooling is headed.
The trade-off is size. A full-text bundle for a large API can run to hundreds of thousands of tokens, and an agent budgeting its context window may decline to load it or truncate it unpredictably. Producers get the most from llms-full.txt when the underlying docs are token-efficient: concise pages, answer-first structure, and markdown rather than rendered HTML. The file rewards documentation that was already written with a machine reader in mind.
Discry's llms-full.txt discovery check fetches /llms-full.txt and passes on presence of a non-empty file, reporting the file's token count in the check detail — so agent builders can see, before loading it, what a full ingest of the docs surface costs.
Checks: llms-full.txt · how each is scored
What does an AI agent make of your API?
Find out in about a minute — no signup.
Discry your API — free