MCP-Bridge Overview

Title

MCP-Bridge – A bridge between the OpenAI API and MCP tools, enabling seamless integration for developers.

How to Use

MCP-Bridge facilitates interaction with MCP tools via the OpenAI API interface.

Installation

Docker (Recommended)

1. Clone the repository.
2. Configure compose.yml with config.json (via volume mount, HTTP URL, or direct JSON).
3. Run:

   docker-compose up --build -d

Manual Installation

1. Clone the repository.
2. Install dependencies:

   uv sync

3. Create a config.json file.
4. Run:

   uv run mcp_bridge/main.py

Usage

• Access API docs at http://yourserver:8000/docs.
• Use OpenAI-compatible endpoints to interact with MCP tools.
• For SSE (Server-Sent Events) support, connect clients to:

  http://yourserver:8000/mcp-server/sse
  

Key Features

• OpenAI API Compatibility – Use any OpenAI API client with MCP tools.
• MCP Tool Integration – Supports MCP tools, sampling, and resources (planned).
• Streaming & Non-Streaming – Works with both streaming and non-streaming completions.
• SSE Bridge – Enables external MCP clients to connect via Server-Sent Events.
• API Key Authentication – Secure your server with Bearer token authentication.

Use Cases

• Open Web UI Integration – Use MCP tools in Open Web UI without explicit MCP support.
• Testing & Debugging – Verify MCP tool configurations using mcp-cli or other clients.
• Custom AI Workflows – Extend OpenAI API functionality with MCP tool calls.
• Secure API Access – Protect your MCP-Bridge instance with API key authentication.

For support, visit the Discord or check the documentation.