NeuronDB

소개

NeuronDB is a comprehensive ecosystem that transforms PostgreSQL into an advanced AI database. It integrates vector search, over 50 machine learning algorithms, and embedding generation directly within the database. The system is modular, comprising NeuronDB (the PostgreSQL extension), NeuronAgent (an agent runtime with REST API and WebSocket support for building AI applications), and NeuronMCP (a Model Context Protocol server for MCP-compatible clients like Claude Desktop). This architecture allows services to operate independently while leveraging a shared PostgreSQL instance, providing a robust platform for developing AI-driven applications with capabilities like hybrid search, RAG pipelines, and GPU acceleration.

주요 기능

• Retrieval-Augmented Generation (RAG) pipeline with LLM integration
• Vector search with HNSW and IVF indexing for similarity search
• Over 50 machine learning algorithms (classification, regression, clustering)
• GPU acceleration for vector and ML operations (CUDA, ROCm, Metal)
• Agent runtime system with REST API, WebSocket, and PostgreSQL-based job queue
• 0 GitHub stars

사용 사례

• Building semantic search and similarity matching applications within PostgreSQL
• Creating autonomous agent systems with long-term memory and tool execution via NeuronAgent
• Developing retrieval-augmented generation (RAG) pipelines for AI models