# Docs for LLMs

Machine-readable exports of these docs: a Markdown twin of every page, the site index at llms.txt, and the Copy page button, for agents and retrieval.

## A Markdown twin of every page

Every docs page is also served as plain Markdown at the same URL with a `.md` suffix. The response is `text/markdown` and starts with the page title as an H1, then the description, then the body.

```bash
curl -s https://datapace.ai/docs/concepts/the-graph.md
```

The Markdown carries a `Link` header pointing at the HTML page as its canonical URL, so an agent that cites the Markdown can hand a person the readable page.

## The site index

[llms.txt](/llms.txt) at the root of the site lists every docs page with its Markdown URL and a one-line description, in sidebar order, alongside the use cases, the comparisons, and the blog. It follows the llms.txt convention: one heading per section, one line per page.

```bash
curl -s https://datapace.ai/llms.txt
```

## Copy page

Each page has a **Copy page** button beside its title. It puts the same Markdown on the clipboard, ready to paste into a chat, an issue, or a prompt. The **View as Markdown** link beside it opens the `.md` twin.

## Citing a page

Cite the HTML URL, not the `.md` twin, so a reader lands on the page with its navigation. The heading ids are stable and derived from the heading text, so a section can be cited directly:

```text
https://datapace.ai/docs/concepts/agents#what-a-proposal-contains
```

## What the exports do not contain

The exports carry the same content as the pages, nothing more. They do not describe how Datapace connects, where it runs, or how data is handled; those are decided with each partner and documented as each surface ships.