소개
This project provides a prototype implementation of an MCP server designed to enable generative AI applications, such as Claude Desktop, to interact with remote backend systems like Apache OFBiz and Moqui. It acts as an intermediary, receiving requests from an MCP client, exposing specific tools, and forwarding those requests to remote systems via RESTful API endpoints. Implemented in TypeScript using the Anthropic SDK, it supports both local (stdio transport) and remote (Streamable HTTP transport) server configurations, allowing AI applications to seamlessly query and retrieve information from these enterprise systems.
주요 기능
- 1 GitHub stars
- Configurable backend API base URL, authentication, and server port
- Provides an example tool for retrieving product information from Apache OFBiz
- Modular design with dynamic discovery of MCP tools from a dedicated directory
- Supports both local (stdio) and remote (Streamable HTTP) MCP server modes
- Includes a script for easily generating and setting OFBiz API authorization tokens
사용 사례
- Enable generative AI to ask clarifying questions (e.g., for a product ID) before executing a backend API call.
- Retrieve specific product information from an Apache OFBiz instance using natural language prompts in generative AI.
- Allow generative AI to compare information from multiple product queries to provide consolidated insights.
Sponsored