This is a pure Java implementation of a connector for the MCP (Model Context Protocol. It acts as a bridge between any MCP-compatible client (like Claude or OpenAI tools) and a standard HTTP backend (like a Spring Boot or Node.js server). You can use with a2ajava which is pure java implementation of A2A protocol and works with MCP protocol as well .

The connector acts as a Passthrough Server. It reads JSON-RPC 2.0 messages from stdin, forwards them as HTTP POSTs to a specified backend URL, and returns the HTTP response as the output to stdout.

1. MCP Tool sends a JSON-RPC message to the connector (via stdin).
2. Connector receives and parses the input.
3. It sends the message as an HTTP POST to your backend (e.g., http://localhost:8080).
4. The backend returns a JSON-RPC-compliant response.
5. Connector prints the response to stdout for the AI tool to consume.

This allows any language backend (Spring, Flask, Express, etc.) to seamlessly speak MCP.

Update your Claude Desktop tool configuration to use this Java connector:

"routeplanningagent": {
  "command": "java",
  "args": [
    "-jar",
    "/work/a2a-mcp-bridge/target/mcp-connector-full.jar",
    "http://localhost:8080"
  ],
  "timeout": 30000
}

or with username

,"laundry": {
  "command": "java",
  "args": [    
    "-jar",
    "/work/a2a-mcp-bridge/target/mcp-connector-full.jar",
    "http://localhost:7860/" ,"user","password"
  ],
  "timeout": 30000
} 

For other types of connectors in python, shell , nodejs please read the guide

• ✅ Language Independence You don’t need to implement MCP in your backend; just handle HTTP POSTs.

• ✅ Better Portability Can be bundled into a single JAR with all dependencies. Run it on any system with Java.

• ✅ Production Ready Java is battle-tested for high performance and robust networking.

• ✅ No NodeJS or Python Needed Ideal for environments with strict runtime policies (e.g., enterprise setups).

• ✅ Custom Logic Support Easily add authentication, logging, metrics, or normalization of payloads.

🏗️ Build Instructions Build the connector with Maven:

mvn clean package

This generates a fat JAR:

💡 Advanced Features Normalizer Support: (Optional) You can add a RequestNormalizer class to rewrite or preprocess requests before they reach the backend.

Debug Logging: All input/output is logged to stderr for easy debugging.