01AI-native graph database with full ISO GQL query engine and optimizer
02Built-in vector search (HNSW), graph-native RAG, and agent memory with semantic search
03Integrated 64-tool Model Context Protocol (MCP) server for AI agent interaction
04Pure Rust, single ~14 MB binary with zero C/C++ dependencies, sub-second cold start, and edge-first deployment options
051 GitHub stars
06On-device embedding inference via EmbeddingGemma (candle), with GPU acceleration (Metal, CUDA)