# MCP Server Copilot

[![MCP Server](https://img.shields.io/badge/MCP-Server-D2C3A5?style=flat-square)](https://modelcontextprotocol.io)
[![python](https://img.shields.io/badge/Python-3.10_%7C_3.11_%7C_3.12-blue?logo=python&logoColor=white&style=flat-square)](https://github.com/tshu-w/mcp-copilot)
[![PyPI - Version](https://img.shields.io/pypi/v/mcp-server-copilot?style=flat-square)](https://pypi.org/project/mcp-server-copilot)
[![uv](https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/astral-sh/uv/main/assets/badge/v0.json&style=flat-square)](https://github.com/astral-sh/uv)
[![Ruff](https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/astral-sh/ruff/main/assets/badge/v2.json&style=flat-square)](https://github.com/astral-sh/ruff)

A meta Model Context Protocol (MCP) server that seamlessly scales LLMs to 1000+ MCP servers through automatic routing without expose all servers and tools to LLMs directly.

<p align="center">
  <img src="https://github.com/user-attachments/assets/7bb3ac79-6706-4fc5-8dab-2f0a3e5b981c" alt="mcp_copilot" style="max-width:100%; height:auto; width:500px;">
</p>

## Components

### Tools

- `router-servers`: Route user query to appropriate MCP servers.
  - `query` (string, required): User's query to find relevant servers.
  - `top_k` (integer, optional): Maximum number of servers to return (default: 5).

- `route-tools`: Route user query to appropriate MCP tools across all servers.
  - `query` (string, required): User's query to find relevant tools.
  - `top_k` (integer, optional): Maximum number of tools to return (default: 5).

- `execute-tool`: Execute a specific tool on a specific server based on previous routing results.
  - `server_name` (string, required): Name of the server hosting the tool.
  - `tool_name` (string, required): Name of the tool to execute.
  - `params` (object, optional): Parameters to pass to the tool, as a key-value dictionary (default: null or empty object).

## Installation

### Using uv (recommended)

When using [`uv`](https://docs.astral.sh/uv/) no specific installation is needed. We will
use [`uvx`](https://docs.astral.sh/uv/guides/tools/) to directly run *mcp-server-copilot*.

### Using PIP

Alternatively you can install `mcp-server-copilot` via pip:

```
pip install mcp-server-copilot
```

After installation, you can run it as a script using:

```
python -m mcp_server_copilot
```

## Configuration

Copy `config/config.sample.json` to `~/.config/mcp-server-copilot`

Add to your MCP Client settings:

<details>
<summary>Using uvx</summary>

```json
{
  "mcpServers": {
    "copilot": {
      "command": "uvx",
      "args": ["mcp-server-copilot", "--config", "~/.config/mcp-server-copilot/config.json"]
    }
  }
}
```
</details>

<details>
<summary>Using pip installation</summary>

```json
{
  "mcpServers": {
    "copilot": {
      "command": "python",
      "args": ["-m", "mcp_server_copilot", "--config", "~/.config/mcp-server-copilot/config.json"]
    }
  }
}
```
</details>

## TODOs

- [ ] Add Dockerfile
- [ ] Tune prompt and add show cases
- [ ] Manage Servers more easily
- [ ] Add Semantic Routing
- [ ] Add Planning Capabilities
- [ ] Resources & Prompts