MCProxy - MCP Proxy for RESTful APIs

MCProxy easley and almost seamlessly generates and deploys a proxy between your existing API and the clients.
It won't replace the existing interface with the existing API, but it will extend the functionality of your system with MCP.

What's MCP?

MCP (Model Context Protocol) is a standardized communication protocol that enables AI tools and services to interact seamlessly, where servers expose operations (“tools”) with defined input/output schemas and clients can discover, invoke, and use them dynamically.
In this ecosystem, an MCP Server acts as a gateway, exposing AI capabilities, workflows, or APIs—often with streaming outputs—while an MCP Client, typically an AI agent or application, discovers and calls these tools to perform tasks.

MCProxy acts as an MCP Server proxy layer that wraps your existing REST API, enabling AI clients to interact with your services using the MCP protocol without changing your underlying API.

Deployment

Deploying MCProxy is straightforward:

1. Build the Docker image:
   docker build -t mcp-proxy:<version> .

2. Prepare the configuration file
   Define the endpoints you want to expose to MCP clients. By default, MCProxy will look for config.json.

3. Mount the configuration file
   Mount it into your Docker container, Kubernetes pod, or deployment.

4. Expose MCProxy
   Make the service accessible via your preferred routing or ingress method.

💡 You can check the examples in ./dev/ for inspiration on configuration and setup.

Once deployed, MCP clients can connect to MCProxy and access exactly the data you’ve allowed — no extra coding required.

Environment Variables

Configuration JSON Schema:

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "title": "API Endpoints Schema",
  "type": "object",
  "required": [
    "version",
    "endpoints"
  ],
  "properties": {
    "version": {
      "type": "string",
      "description": "Version of the schema or API definition"
    },
    "endpoints": {
      "type": "array",
      "description": "List of API endpoints",
      "items": {
        "type": "object",
        "required": [
          "name",
          "description",
          "url",
          "method"
        ],
        "properties": {
          "name": {
            "type": "string",
            "description": "Unique name of the endpoint"
          },
          "description": {
            "type": "string",
            "description": "A short description of the endpoint"
          },
          "url": {
            "type": "string",
            "format": "uri",
            "description": "Base URL of the endpoint"
          },
          "method": {
            "type": "string",
            "enum": [
              "GET",
              "POST",
              "PUT",
              "DELETE",
              "PATCH"
            ],
            "description": "HTTP method to use"
          },
          "path_params": {
            "type": "array",
            "description": "List of path parameters",
            "items": {
              "type": "object",
              "additionalProperties": {
                "type": "string"
              }
            }
          },
          "query_params": {
            "type": "array",
            "description": "List of query parameters",
            "items": {
              "type": "object",
              "additionalProperties": {
                "type": "string"
              }
            }
          }
        },
        "additionalProperties": false
      }
    }
  },
  "additionalProperties": false
}

Example for a configuration file: ./dev/deployment/docker/endpoints.json

Testing

Set up the environment, including:

• Nginx as the reverse proxy (BSD 2-Clause License)
• mendhak/http-https-echo as the RESTful API (MIT License)
• The MCProxy server

cd ./dev/deployment/docker
docker compose up -d

To test the default configuration, you can use the available scripts, for example:

cd ./dev/scripts
python3 test_list_tools.py

Note: Both Nginx and http-https-echo are used only in the example Docker Compose setup and are not bundled with this project.

⚠️ Disclaimer:

The creators of MCProxy are not responsible for any unintended exposure of data through the endpoints you configure. It is your responsibility to ensure that the endpoints you expose do not leak sensitive or unwanted information.