AI Memory & Knowledge Engine
15.2k 2026-04-13
topoteretes/cognee
An open-source knowledge engine designed to build personalized and dynamic memory for AI agents with minimal code.
Core Features
Unified data ingestion (any format or structure)
Combines vector search, graph databases, and cognitive science approaches
Continuously learns and provides relevant context for AI agents
Supports persistent, learning, and trustworthy agents
Runs locally, offers ontology grounding, and multimodal capabilities
Detailed Introduction
Cognee is an open-source knowledge engine that empowers developers to create sophisticated memory systems for AI agents. It facilitates the ingestion of diverse data formats and structures, continuously learning to provide precise context. By integrating vector search, graph databases, and cognitive science principles, Cognee makes documents searchable by meaning and interconnected by relationships, ensuring agents have dynamic, evolving knowledge. It aims to build reliable, persistent, and learning AI agents.