Cognee provides a compact Python API to ingest documents, build knowledge graphs, and run graph-aware searches so agents can store and reuse long-term and session-based memories. It supports ingestion (add), graph construction (cognify), enrichment (memify), and multiple search modes (graph, temporal, RAG, summaries).
Use Cognee when you need agent memory that persists across sessions, when you want graph-structured retrieval (multi-hop or temporal queries), or when you need scoped memories per user/project via datasets and NodeSets. Good for personalization, feedback loops, and improving agent behaviour over time.
add -> cognify -> search workflow, guidance on DataPoint models, NodeSets, and SearchType selection.Best used with Python-based agents and frameworks that can call async Python APIs (e.g., Codex/Copilot-style integrations, Python-backed agent runtimes).
This skill has not been reviewed by our automated audit pipeline yet.
