Spring AI MCP PoC

What is Spring AI MCP PoC?

Spring AI MCP PoC is a proof-of-concept application that integrates the Spring AI framework with GitHub’s Model Context Protocol (MCP) server and OpenAI’s GPT models. It processes natural language prompts, enables function calling for GitHub operations, and provides a RESTful API for AI-powered interactions.

How to use Spring AI MCP PoC?

Clone the repository, set environment variables (OPENAI_API_KEY required; GITHUB_MCP_SERVER_URL and GITHUB_TOKEN optional), then build and run with ./gradlew build and ./gradlew bootRun. Use POST /api/v1/prompt to process prompts (JSON with "prompt" and "enableMcpTools" fields) and GET /api/v1/mcp/tools to list available MCP tools.

Key features of Spring AI MCP PoC

• Spring AI integration for LLM interactions
• GitHub MCP support via Model Context Protocol
• AI function calling for GitHub operations
• RESTful API endpoints for prompt processing
• Configurable environment variables and application properties
• Robust error handling and logging

Use cases of Spring AI MCP PoC

• Build an AI assistant that queries and manages GitHub repositories
• Prototype a natural‑language interface for GitHub APIs
• Test and evaluate MCP integration with Spring AI and OpenAI

FAQ from Spring AI MCP PoC

What are the prerequisites to run Spring AI MCP PoC?

You need Java 17+, Gradle 8.6+, and an OpenAI API key. Optionally, a running GitHub MCP server and a GitHub personal access token for enhanced features.

How do I configure the GitHub MCP server?

Set the GITHUB_MCP_SERVER_URL environment variable or the app.github.mcp.server.url property in application.yml. The server must provide repository, issue, pull request, and code search endpoints.

What happens if no GitHub MCP server is available?

The application still works but with limited GitHub integration. Basic prompt processing and AI responses will function without the MCP server.

Which API endpoints are available?

GET /api/v1/health – health check, POST /api/v1/prompt – process an AI prompt, and GET /api/v1/mcp/tools – list available MCP tools.

What technologies does Spring AI MCP PoC use?

Spring Boot 3.2.5, Spring AI 1.0.0-M1, OpenAI GPT-4, WebFlux, Jackson, and Gradle 8.6.