llms.txt vs llms-full.txt
Same idea, two sizes. llms.txt is a curated index — a short Markdown list of your best pages with one-line descriptions. llms-full.txt takes those same pages and inlines their full text into a single file. One is a table of contents; the other is the whole book bound into one document.
| llms.txt | llms-full.txt | |
|---|---|---|
| Contains | Links + short descriptions | The full text of those pages |
| Size | Small — a curated index | Large — everything inlined |
| Best for | Navigation & prioritization | RAG, chatbots, one-shot context |
| Fetches needed | AI follows links to read more | Zero — it's all in one file |
| Format | Markdown (headings, links) | Markdown (headings + body text) |
| Location | /llms.txt | /llms-full.txt |
| Analogy | Table of contents | The whole book in one file |
What each file looks like
llms.txt — the curated index
# Example Co > Open-source analytics for indie developers. ## Docs - [Quickstart](https://example.com/docs/quickstart): install in 5 minutes - [API reference](https://example.com/docs/api): every endpoint ## Optional - [Changelog](https://example.com/changelog)
llms-full.txt — the same pages, full text inlined
# Example Co > Open-source analytics for indie developers. ## Quickstart Install the CLI with `npm i -g example`. Run `example init` in your project root to create a config file. Then `example start` boots the local dashboard on port 4000. Full walkthrough: 1. Install the package globally... (...the entire Quickstart page, inlined...) ## API reference ### GET /events Returns a paginated list of events. Query params: - `from` (ISO date) — start of range (...the entire API page, inlined...)
When to use which
- Reach for llms.txt when you want an AI (or a person) to see your structure and decide what's worth reading — a lightweight map that keeps its own footprint tiny.
- Reach for llms-full.txt when you want the content available in one shot: feeding a RAG index, a docs-search embedding job, or pasting a whole product's docs into a chat window with zero follow-up fetches.
They're not either/or. Publish llms.txt as the index and llms-full.txt as the full-text companion — the map and the territory, both at your site root.
Generate both files in one click — free, no signup. We're one of the few tools that outputs llms-full.txt too.
Generate llms.txt + llms-full.txt →FAQ
What's the difference between llms.txt and llms-full.txt?
llms.txt is a short, curated index — a Markdown list of your key pages with one-line descriptions and links. llms-full.txt is the opposite: it inlines the actual content of those pages into one large file, so an AI has the full text in a single request instead of a set of links to follow.
Do I need both llms.txt and llms-full.txt?
They serve different jobs, so most sites publish both. llms.txt helps an AI (or a person) understand your structure and pick what to read; llms-full.txt gives it the complete content immediately without extra fetches. Publishing both covers the map-first and content-first cases.
Which file should I use for a RAG pipeline or chatbot?
llms-full.txt. Because it's a single file with the full text, you can drop it straight into a retrieval index, a docs-search embedding job, or a context window — no crawling, no per-page requests. llms.txt is better as a lightweight table of contents.
Where do llms.txt and llms-full.txt live?
Both sit at your site root: /llms.txt and /llms-full.txt. The naming and locations follow the llms.txt proposal from Answer.AI (2024). Neither replaces robots.txt or your sitemap.
Is llms-full.txt part of the official spec?
The original llms.txt proposal describes llms.txt as the curated index and notes that expanded, full-content variants (commonly named llms-full.txt) are a natural companion. Treat llms.txt as the core file and llms-full.txt as the widely-adopted full-text sibling.
Related: What is llms.txt? · Is llms.txt worth it? · llms.txt vs robots.txt