# llms.txt & Markdown Docs

Feed these docs to AI assistants with llms.txt, Markdown versions of every page, and one-click copy and open actions.


Language models read Markdown far better than rendered HTML with navigation, scripts and styling around it. So every page of these docs is also available as clean Markdown, and an `llms.txt` file indexes them all.

## llms.txt

[https://social-publish-kappa.vercel.app/llms.txt](https://social-publish-kappa.vercel.app/llms.txt) follows the [llms.txt](https://llmstxt.org) convention: a plain-text Markdown index of the whole documentation, with a link and short description for every page. Give it to an assistant as a starting point and it can find and fetch the pages it needs.

```text
https://social-publish-kappa.vercel.app/llms.txt
```

Use it to:

- paste into a chat with "Use these docs to help me integrate Relaycast",
- point an agent with web access at the docs,
- reference it from an [agent skill](/docs/ai/agent-skill) or rules file so coding agents can look up details.

## Markdown for every page

Add `.md` to any docs URL to get that page as Markdown:

| Page | HTML | Markdown |
| --- | --- | --- |
| Getting Started | `https://social-publish-kappa.vercel.app/docs` | `https://social-publish-kappa.vercel.app/docs.md` |
| Authentication | `https://social-publish-kappa.vercel.app/docs/authentication` | `https://social-publish-kappa.vercel.app/docs/authentication.md` |
| Platform guide | `https://social-publish-kappa.vercel.app/docs/platforms/tiktok` | `https://social-publish-kappa.vercel.app/docs/platforms/tiktok.md` |
| API reference | `https://social-publish-kappa.vercel.app/docs/api/create-post` | `https://social-publish-kappa.vercel.app/docs/api/create-post.md` |

The pattern is `https://social-publish-kappa.vercel.app/docs/<slug>.md`, and the Getting Started page, which lives at `/docs`, is `/docs.md`.

The Markdown has the page title and description at the top, and the same content as the HTML page: headings, tables, callouts and code samples, with real API URLs filled in. Generated sections, such as the option tables on platform pages, are included too.

Fetch it from a script like any other text file:

```bash
curl https://social-publish-kappa.vercel.app/docs/scheduling.md
```

## Copy and open from any page

Every docs page has two actions next to the title:

- **Copy Markdown** copies the page's Markdown to your clipboard, ready to paste into any assistant.
- **Open in ChatGPT / Claude** opens a new chat with that assistant, preloaded with a prompt that points it at the page so you can ask questions about it straight away.

These are the fastest way to get help with one specific topic, for example pasting the [TikTok guide](/docs/platforms/tiktok) before asking why your TikTok posts fail validation.

## Tips for using docs with assistants

- **Give the right page, not all of them.** A focused page (for example [Posts](/docs/posts) and the guide for your platform) usually gets better answers than the whole site. Use `llms.txt` when you want the assistant to choose.
- **Include your request and the error.** Paste the request body and the full error JSON, including `details`. The [Playground](/docs/playground) can generate the request for you.
- **Never paste your access key.** Replace it with a placeholder such as `$RELAYCAST_API_KEY` before sharing code or cURL commands.
- **Check generated code against the reference.** Assistants sometimes invent endpoints or SDKs. Relaycast has no official SDK, and the public API is the endpoints listed in the [API Reference](/docs/api).