# Use these docs with LLMs and AI agents

> Give an AI assistant every Coritan guide as plain Markdown, through llms.txt, llms-full.txt, a page's index.md or the search API.

Source: https://docs.coritan.com/get-started/for-llms/

Every page on this site is also plain Markdown, so a large language model (LLM) or an AI agent can read the same guides you do, without the site's layout around them. Point your assistant at one of the files below, and it can answer questions about Coritan and walk you through a task with the exact labels and API calls.

## The files

[`/llms.txt`](/llms.txt)
: The index. It links every guide on this site, section by section, each with one line on what it covers, and names the files below. Give an agent this address first: it can then fetch only the pages it needs.

[`/llms-full.txt`](/llms-full.txt)
: Every guide in one file, in the order the sidebar shows them. It is large (more than 2 MB), so use it with an assistant that can take a long context, or pick a section's file instead.

`/<section>/llms-full.txt`
: One section in one file, such as [`/mail/llms-full.txt`](/mail/llms-full.txt) or [`/billing/llms-full.txt`](/billing/llms-full.txt). The section's slug is the first part of its pages' addresses.

[`/api/llms-full.txt`](/api/llms-full.txt)
: The API guides and the whole generated API reference: every Client API and Organization API operation with its parameters, request body and answers. `/llms-full.txt` leaves the reference out, because it is larger than every guide together.

`<page>/index.md`
: Any single page as Markdown. Add `index.md` to the page's address: [/get-started/for-llms/index.md](/get-started/for-llms/index.md) is this page.

`/search.json?q=`
: The site's search, as JSON. `/search.json?q=floating+ip` answers `{"query": ..., "results": [...]}` with up to 12 results, each with its `title`, `url`, `section`, `kind` and a `snippet`.

## Give the docs to an assistant

*An agent that can browse*
: Such as a coding agent or an assistant with web access. Tell it to read `https://docs.coritan.com/llms.txt` and follow the links it needs.

*A chat assistant without browsing*
: Download [`/llms-full.txt`](/llms-full.txt), or one section's file, and attach it to the conversation.

*Your own tool or retrieval index*
: Fetch `/llms.txt`, then each page's `index.md`, and index those. The [sitemap](/sitemap.xml) lists every page's address too.

The files change whenever the docs do, so fetch them again rather than keeping an old copy. Every change to Coritan that a customer could notice updates these docs at the same time.

## What the Markdown holds

Each page's Markdown starts with its title, its one-line summary and a `Source:` line with the page's address, then has the same steps, tables, examples and links as the page. Links to other pages are site paths such as `/billing/add-credit/`, which an agent opens on docs.coritan.com, with `index.md` added for the Markdown.

An assistant is only as right as what it reads. When an answer matters, such as a price, a limit or a step that deletes data, open the page it cites and check.

## Related

- [Make your first API request](/get-started/first-steps-with-the-api/)
- [API](/api/)