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

# MCP server

> Model Context Protocol server for integrating OpenBunny with LLM tools.

The service includes an MCP (Model Context Protocol) server that exposes task management as tools for LLM applications like Claude Desktop, Cursor, or any MCP-compatible client.

## Available tools

| Tool              | Description                                                |
| ----------------- | ---------------------------------------------------------- |
| `list_tasks`      | List tasks with optional status and priority filters       |
| `get_task`        | Fetch a single task by ID                                  |
| `update_task`     | Modify task fields (status, priority, labels, description) |
| `delete_task`     | Remove a task                                              |
| `search`          | Hybrid search across tasks                                 |
| `find_duplicates` | Find tasks similar to a given text                         |
| `set_task_status` | Change a task's status                                     |

## Connecting

The MCP server runs on the same port as the REST API (default: 3100). Connect using the standard MCP transport:

```bash theme={null}
# MCP endpoint
POST http://localhost:3100/mcp
GET  http://localhost:3100/mcp   # SSE transport
```

### Claude Desktop configuration

Add OpenBunny to your Claude Desktop MCP configuration:

```json claude_desktop_config.json theme={null}
{
  "mcpServers": {
    "openbunny": {
      "url": "http://localhost:3100/mcp"
    }
  }
}
```

Once connected, you can ask Claude to manage your tasks directly: "Show me all high-priority tasks" or "Create a task to fix the login bug."
