MCPizer

MCPizer is an MCP server that converts REST APIs, OpenAPI definitions, and gRPC or Connect-RPC services into Model Context Protocol tools for AI assistants. It connects directly to backend services running frameworks like FastAPI, Spring Boot, or Express, as well as standalone gRPC endpoints using server reflection or direct Protocol Buffer specifications. Developers and engineers use MCPizer to let LLM clients such as Claude Desktop or VS Code invoke internal APIs, query databases via REST endpoints, and run microservice remote procedure calls without writing custom MCP wrappers for every endpoint. The server automatically discovers OpenAPI paths from base URLs, parses raw schemas hosted locally or on GitHub, and binds service methods to schema-validated tool definitions. When an AI client executes a tool, MCPizer handles request serialization, issues the network call using either standard HTTP or gRPC/Connect protocols, and formats the response for the model. It operates in both STDIO mode for local desktop integration and Server-Sent Events mode over HTTP for remote client access.

Category: AI & LLM Tooling

Tags: automation, grpc, integration, openapi, rest

Visit MCPizer

How to install and configure MCPizer

1. Install the MCPizer binary using the Go toolchain: bash go install github.com/i2y/mcpizer/cmd/mcpizer@latest Ensure that $GOPATH/bin is in your system PATH. 2. Create a configuration file named config.yaml specifying your API schemas or service URLs: yaml schema_sources: - https://api.example.com/openapi.json - grpc://localhost:50051 3. Register MCPizer in your Claude Desktop configuration file (claude_desktop_config.json): json { "mcpServers": { "mcpizer": { "command": "mcpizer", "args": ["-transport=stdio", "-config=/path/to/config.yaml"] } } } 4. Restart Claude Desktop or your MCP client to load the generated tools.

What you can do with MCPizer

• Exposing FastAPI or Spring Boot service endpoints directly to Claude Desktop by auto-discovering their OpenAPI documentation routes. - Invoking backend gRPC methods from an AI assistant using protobuf files or runtime gRPC server reflection without custom wrapper code. - Interacting with Connect-RPC microservices in HTTP/JSON mode to inspect payloads and test distributed services via natural language prompts. - Querying internal private company APIs using GitHub-hosted OpenAPI specifications authenticated via the GitHub command-line interface tool. - Serving unified MCP tools across local development environments and remote production servers using separate configuration targets and endpoints.

Key facts

• https://github.com/i2y/mcpizer
• AI & LLM Tooling, Developer Tools & Code Intelligence
• automation, grpc, integration, openapi, rest

Part of MCP Servers

How do I install MCPizer?

You can install MCPizer using the Go toolchain by running go install github.com/i2y/mcpizer/cmd/mcpizer@latest. Ensure that your GOPATH bin directory is included in your system PATH environment variable so the executable can be run directly from the terminal or launched by desktop MCP clients.

What can MCPizer do?

MCPizer dynamically transforms REST APIs and gRPC services into Model Context Protocol tools. It auto-discovers OpenAPI schemas from running web applications, reads schema files from local paths or GitHub repositories, supports Connect-RPC protocols, and manages tool invocation requests from AI assistants directly to target service endpoints.

Which MCP clients work with MCPizer?

MCPizer works with any MCP-compliant client. It supports standard input and output communication for clients that manage local subprocesses, such as Claude Desktop and VS Code extensions. It also provides a Server-Sent Events mode over HTTP, allowing external or network-connected clients to access tools at a local or remote endpoint.

Does MCPizer support private GitHub repositories?

Yes. MCPizer supports schema paths starting with the github:// protocol prefix. When accessing private repositories, it automatically authenticates requests using your locally configured GitHub CLI credentials, allowing you to load private OpenAPI definitions or Protocol Buffer files without placing API keys in your configuration files.

Can MCPizer call gRPC services without reflection?

Yes. While MCPizer can discover schemas from gRPC reflection if enabled, it also accepts direct links to proto files hosted locally, via HTTP, or in GitHub repositories. You can specify the proto file location along with the target gRPC server address in your configuration file.