mcp-bridgekit

MCP BridgeKit is a Python-based HTTP gateway that exposes MCP stdio servers over HTTP. It runs one or more per-user MCP sessions in a pool, translating incoming HTTP requests into JSON-RPC calls to your MCP servers, and streaming results back to clients. The server supports per-user session pooling (up to 100 concurrent sessions), background job handling for long-running calls, and live discovery of available tools. You can interact with it via standard HTTP endpoints to initiate chats, list available tools for a user, and monitor background jobs.

To use it, deploy the Python server (for example via a container or your preferred hosting platform) and point your web app to its HTTP API. With a POST to /chat you provide details such as user_id and the MCP tool you want to call along with arguments. The gateway ensures each user gets an isolated MCP session, so tool calls and state do not leak across users. If a call might exceed typical HTTP timeouts, BridgeKit can queue the work in Redis and provide a job_id for polling status and results, making it suitable for web apps with shorter request time limits.

Prerequisites:

• Python 3.11+ installed on the host
• Redis server accessible to BridgeKit for session and job storage
• Internet access for installing Python packages

Step-by-step installation:

1. Create and activate a Python virtual environment (recommended): python -m venv venv source venv/bin/activate # on Unix/macOS venv\Scripts\activate # on Windows

2. Install the BridgeKit package from PyPI (example package name assumed as mcp-bridgekit): pip install mcp-bridgekit

3. Configure environment variables (example): export MCP_BRIDGEKIT_REDIS_URL=redis://user:pass@host:6379/0

4. Run the BridgeKit server (as described in the README): python -m mcp_bridgekit

5. Verify the server is running by sending a test request or visiting the health endpoint if provided in the docs.

Environment Variables:

• MCP_BRIDGEKIT_REDIS_URL: Redis URL used for session pooling and background job storage. Ensure Redis is reachable from your deployment environment.
• Depending on deployment, you may configure additional logging or diagnostics via standard Python logging configuration or environment variables described in the project docs.

Common issues:

• Redis connectivity failures: verify host, port, and credentials; ensure network access from the BridgeKit host.
• Exceeding the per-user session pool limit: monitor active sessions and adjust pool size or client request patterns.
• Long-running tool calls: use the async/queue mechanism and poll /job/{id} to avoid HTTP timeouts.