Show HN: AI-powered web service combining FastAPI, Pydantic-AI, and MCP servers
a day ago
- #AI
- #FastAPI
- #MCP
- A production-ready AI-powered web service combining FastAPI, Pydantic-AI, and MCP servers.
- Features include high-performance asynchronous APIs with FastAPI, type-safe agent construction with Pydantic-AI, and plug-and-play tools via MCP servers.
- The service provides a UI for querying trends from Hacker News and Web Search, delivering ranked trend cards with scores, links, and AI-written summaries.
- Includes advanced data modeling with Pydantic, multi-agent AI systems with A2A communication, and Docker containerization for AI services.
- Deployment options include Docker Compose for easy setup, with environment variables for API keys and tokens.
- Endpoints available for trends analysis, repository insights, combined analysis, and general assistant queries.
- Agents include Entry Agent for tech trend analysis and Specialist Agent for GitHub repository intelligence, communicating via A2A protocol.
- MCP servers like Brave Search and Hacker News provide structured data inputs, with optional GitHub and Filesystem MCP servers.
- Development tools include pytest for testing, mypy for type checking, and Black/Ruff for code formatting and linting.
- Project structure is organized with clear separation of agents, models, services, and utilities, supporting extensibility and customization.