> ## Documentation Index
> Fetch the complete documentation index at: https://docs.nowrun.io/llms.txt
> Use this file to discover all available pages before exploring further.

# For AI Agents: MCP Server, llms.txt, and Skills for nowrun Docs

> Endpoints and files that expose the nowrun documentation to AI coding agents — llms.txt, /mcp, skill.md, and MCP discovery — so agents can search, read, and act on the docs without extra configuration.

nowrun is designed to be driven by AI coding agents. This page lists every agent-facing endpoint the docs site exposes so a cold agent can discover, search, and act on the docs without any manual setup.

## Content indexes

<Card title="/llms.txt" icon="file-text" href="https://docs.nowrun.io/llms.txt">
  A concise index of every page in the docs, formatted for LLMs. Follows the [llmstxt.org](https://llmstxt.org) convention.
</Card>

<Card title="/llms-full.txt" icon="file-lines" href="https://docs.nowrun.io/llms-full.txt">
  The full contents of the docs as a single markdown file. Drop it into an agent's context window when you want everything at once.
</Card>

<Card title="Any page as markdown" icon="code" href="https://docs.nowrun.io/quickstart.md">
  Append `.md` to any docs URL to get the raw markdown for that page (for example, `/quickstart.md`).
</Card>

## MCP server

The docs expose a hosted Search MCP server so agents can search and retrieve pages in-flight instead of relying on stale training data.

| Endpoint                                                   | Purpose                                                                       |
| ---------------------------------------------------------- | ----------------------------------------------------------------------------- |
| `https://docs.nowrun.io/mcp`                               | Search MCP server. Add this URL to any MCP-compatible client.                 |
| `https://docs.nowrun.io/.well-known/mcp`                   | MCP discovery document. Agents check this to locate the server automatically. |
| `https://docs.nowrun.io/.well-known/mcp.json`              | Same discovery document, JSON extension.                                      |
| `https://docs.nowrun.io/.well-known/mcp/server-card.json`  | MCP server card with tool metadata (name, description, input schema).         |
| `https://docs.nowrun.io/.well-known/mcp/server-cards.json` | List of server cards for this deployment.                                     |

Connect from your client of choice:

<CodeGroup>
  ```bash Claude Code theme={null}
  claude mcp add --transport http nowrun https://docs.nowrun.io/mcp
  ```

  ```json Cursor (mcp.json) theme={null}
  {
    "mcpServers": {
      "nowrun": {
        "url": "https://docs.nowrun.io/mcp"
      }
    }
  }
  ```

  ```toml Codex (~/.codex/config.toml) theme={null}
  [mcp_servers.nowrun]
  url = "https://docs.nowrun.io/mcp"
  ```

  ```json VS Code (.vscode/mcp.json) theme={null}
  {
    "servers": {
      "nowrun": {
        "type": "http",
        "url": "https://docs.nowrun.io/mcp"
      }
    }
  }
  ```
</CodeGroup>

## Skills

The nowrun CLI installs a **skill and MCP server** locally when you run `nowrun init`. That skill teaches your agent *how* to use nowrun end-to-end — not just how to search the docs. See [Authentication](/setup/authentication) and [nowrun init](/cli/init).

## Crawler endpoints

| Endpoint                             | Purpose                       |
| ------------------------------------ | ----------------------------- |
| `https://docs.nowrun.io/sitemap.xml` | Full page index for crawlers. |
| `https://docs.nowrun.io/robots.txt`  | Crawler rules.                |

## What's safe to share

All of the endpoints above expose only content that is already publicly published on this site. There are no tokens, secrets, or private data behind them — the same information is reachable by opening the docs in a browser.