> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.caesar.xyz/2025-11-27/documentation/integration/mcp-docs-server/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.caesar.xyz/_mcp/server.

# MCP Docs Server

Caesar's documentation includes a docs MCP (Model Context Protocol) server designed for coding agents building applications that integrate with Caesar. MCP is an [open standard by Anthropic](https://modelcontextprotocol.io/) that enables AI applications to connect with external data sources.

> **Warning**
>
> This MCP server provides access to Caesar's **documentation only**. It helps agents understand how to integrate with Caesar, but it does **not** provide tools to interact with Caesar APIs.

> **Info**
>
> Need MCP tools for the research API? See the [Research MCP Server](/integration/mcp-research).

## Server URL

```
https://docs.caesar.xyz/_mcp/server
```

## Setup

### OpenCode

Add the docs MCP server to your OpenCode config under `mcp` (see the [OpenCode MCP docs](https://opencode.ai/docs/mcp-servers/)):

**`opencode.json`**

```json title="opencode.json"
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "caesar-docs": {
      "type": "remote",
      "url": "https://docs.caesar.xyz/_mcp/server",
      "enabled": true
    }
  }
}
```

### Cursor

Add a remote MCP server in `mcp.json` (see [Cursor MCP docs](https://cursor.com/docs/context/mcp)):

**`mcp.json`**

```json title="mcp.json"
{
  "mcpServers": {
    "caesar-docs": {
      "url": "https://docs.caesar.xyz/_mcp/server"
    }
  }
}
```

### Claude Code

Use the CLI to add the remote HTTP server (see [Claude Code MCP docs](https://code.claude.com/docs/en/mcp)):

```bash
claude mcp add --transport http caesar-docs https://docs.caesar.xyz/_mcp/server
```

### Codex

Add the server in `~/.codex/config.toml` (see [Codex MCP docs](https://developers.openai.com/codex/mcp/)):

**`config.toml`**

```toml title="config.toml"
[mcp_servers.caesar_docs]
url = "https://docs.caesar.xyz/_mcp/server"
```

### Droid

Add the HTTP server with the CLI (see [Droid MCP docs](https://docs.factory.ai/cli/configuration/mcp)):

```bash
droid mcp add caesar-docs https://docs.caesar.xyz/_mcp/server --type http
```

## Usage

Once connected, you can ask your AI assistant questions about Caesar:

```
How do I set up authentication for the Caesar API?
```

```
What's the difference between reasoning_loops and reasoning_mode?
```

```
Show me how to use Caesar with the OpenAI SDK.
```

The MCP server queries Caesar's documentation and returns relevant information directly in your AI tool.

## Available tools

The MCP server exposes a search tool that queries Caesar documentation:

| Tool          | Description                                                |
| ------------- | ---------------------------------------------------------- |
| `search_docs` | Search Caesar documentation using natural language queries |

## Example queries

| Question                                  | What you'll get                                  |
| ----------------------------------------- | ------------------------------------------------ |
| "How do I authenticate?"                  | API key setup, header format, examples           |
| "What parameters control research depth?" | `reasoning_loops`, `reasoning_mode`, `auto` mode |
| "How do I stream results?"                | SSE endpoint, event format, code examples        |
| "What are the rate limits?"               | Concurrency caps, usage limits                   |

## When to use MCP vs llms.txt

| Use case                                          | Recommended                       |
| ------------------------------------------------- | --------------------------------- |
| Quick questions while coding                      | **MCP Docs Server**               |
| Loading full docs into context                    | [llms.txt](/integration/llms-txt) |
| IDE/agent integration (Cursor, Claude Code, etc.) | **MCP Docs Server**               |
| ChatGPT or web-based AI                           | [llms.txt](/integration/llms-txt) |

#### MCP Docs Server

Best for interactive Q\&A with a coding agent. Ask questions, get targeted answers from the docs.

#### [llms.txt](/integration/llms-txt)

Best for bulk context. Load entire documentation sections into your AI's context window.

> Give coding agents targeted access to Caesar documentation