How I Built an Autonomous AI Tool Engine with MCP, Llama 3.3, and Automated Hugging Face Pipelines

AI Tool Hunter V2 reimagines an AI tool directory as an autonomous discovery engine.

A daily GitHub Actions pipeline scans the web and uses Llama 3.3 via Groq to extract metadata and intent-based categories.

The system deduplicates entries, stores them in Supabase, and publishes an updated corpus to Hugging Face.

It is also exposed through a Model Context Protocol server, letting tools like Cursor, VS Code, and Claude Desktop search it directly.