feifei-proxy-mcp

```markdown <div align="center"> <img src="images/proxy-mcp-logo.png" width="40%" /> <h1> MCP Transport Type Conversion Service </h1> </div> <hr/> > This service can convert existing MCP types of stdio / sse to streamable_http transport type MCP. > > This implementation is fully compliant with the official MCP standards. > > Version Update: > > Current Version: 0.0.1 ## Quick Start ### Local Deployment (If local deployment debugging is not needed, skip to containerized deployment) > Python environment requirement: Python 3.12 > > If uv is not installed, please follow the steps below to install it; if it is already installed, you can ignore this. > > uv official website: https://docs.astral.sh/ #### macOS / Linux uv Installation ```shell curl -LsSf https://astral.sh/uv/install.sh | sh ``` #### Windows uv Installation ```shell powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex" ``` <hr/> ## Environment Variable Information | Parameter | Description | Default Value | Required | Notes | |-----------------------------|----------------------------|------------------------|----------|------------------------------------------------| | `TRANSPORT_TYPE` | Transport protocol type | `streamable_http` | No | Fill in the transport protocol type, optional values: `stdio`, `sse`, `streamable_http` | | `PROXY_MCP_NAME` | Name of the proxy MCP server | `feifei-proxy-mcp` | No | Name of the proxy MCP service | | `PROXY_MCP_SERVER_CONFIG` | MCP server configuration | - | Yes | Configuration of the MCP service to be proxied | | `PROXY_MCP_PORT` | Service port | `8000` | No | Used when the proxy MCP protocol type is `sse` or `streamable_http` | <hr/> ## Containerized Deployment > Please ensure that Docker is installed on the current machine before use. ### Docker Command Line Deployment > The following example demonstrates deploying fetch MCP to convert stdio to streamable_http MCP. ```shell docker run -e PROXY_MCP_SERVER_CONFIG='{"mcpServers":{"fetch":{"command":"uvx","args": ["mcp-server-fetch"]}}}' crpi-i8di4kksl3oe9ugm.cn-hangzhou.personal.cr.aliyuncs.com/feifei-ai/feifei-proxy-mcp:latest ``` ### Docker-Compose Deployment (Recommended for quick deployment) > Using docker-compose allows you to deploy multiple proxy MCPs simultaneously. > > For example, in the following example, I deployed three proxy MCPs using docker-compose: > > feifei-proxy-fetch => Web content scraping streamable_http MCP > > feifei-proxy-bazi => Birth chart analysis streamable_http MCP > > feifei-proxy-howtocook => Recipe analysis streamable_http MCP ```yml # docker-compose.yml version: '3.8' services: feifei-proxy-fetch: image: 'crpi-i8di4kksl3oe9ugm.cn-hangzhou.personal.cr.aliyuncs.com/feifei-ai/feifei-proxy-mcp:latest' ports: - "8001:8000" environment: PROXY_MCP_SERVER_CONFIG: '{"mcpServers":{"fetch":{"command":"uvx","args": ["mcp-server-fetch"]}}}' feifei-proxy-bazi: image: 'crpi-i8di4kksl3oe9ugm.cn-hangzhou.personal.cr.aliyuncs.com/feifei-ai/feifei-proxy-mcp:latest' ports: - "8002:8000" environment: PROXY_MCP_SERVER_CONFIG: '{"mcpServers":{"Bazi":{"command":"npx","args":["bazi-mcp"]}}}' feifei-proxy-howtocook: image: 'crpi-i8di4kksl3oe9ugm.cn-hangzhou.personal.cr.aliyuncs.com/feifei-ai/feifei-proxy-mcp:latest' ports: - "8003:8000" environment: PROXY_MCP_SERVER_CONFIG: '{"mcpServers":{"howtocook-mcp":{"command":"npx","args":["-y","howtocook-mcp"]}}}' ``` After completing the docker-compose.yml configuration, execute the deployment command to start the MCP service. ```shell # Start docker compose up -d # After successful deployment, check the container status with docker compose ps ```  <hr/> ## Client Access > Here, we take [cherry stdio](https://www.cherry-ai.com/) as an access example. > If not installed, you can install it yourself or add the MCP server in your own cline, cursor to complete the use.   ```