Cognee
An open-source knowledge graph memory engine designed for AI agents. It provides cross-session persistent long-term memory for agents through a hybrid graph-plus-vector architecture and supports self-hosted deployment.
Timeline (last 90 days)
Cognee is closer to a knowledge graph construction tool, emphasizing the organization of documents and entities with graph structures and providing hybrid graph plus vector retrieval.
What Mem0, Honcho and Cognee have in common is that their memory capabilities are built on general-purpose storage, shielding underlying differences through an abstraction layer.
Cognee, on its improve page, states its gates: guidance is eligible only when it has not been marked harmful and its confidence passes the distillation gate; proposed lessons are handed to the writer/rejecter to check their relationship with existing lessons and graph entities.
The Cognee plugin, after each conversation turn completes, only writes the content to cheap session memory and does not build a graph; graph construction (chunking, LLM extraction of entities and relationships, embedding, and committing edges) is concentrated at the end of the session and performed in a separate process via cognee.improve()
The Cognee local server runs in agent mode and automatically retires within 60 seconds after the last agent deregisters
Cognee integration supports three connection modes: local-server (default), remote (set COGNEE_BASE_URL), and embedded (COGNEE_EMBEDDED=true).
Cognee ingests content and builds a knowledge graph, enabling retrieval to traverse along relationship chains rather than returning isolated text chunks
The Cognee plugin adopts a no-silent-fallback design: if COGNEE_BASE_URL is set but connection fails or the local server fails to start, initialization directly throws an error rather than switching to another mode.
Switching models is fully decoupled from the memory layer, and Cognee treats different models such as Gemini and Opus the same at the underlying level.
To install the Cognee plugin use the command pip install cognee-integration-hermes-agent, and upgrading requires adding the -U parameter.
3 more timeline events
All Facts (13)
Cognee is closer to a knowledge graph construction tool, emphasizing the organization of documents and entities with graph structures and providing hybrid graph plus vector retrieval.
50%UnverifiedWhat Mem0, Honcho and Cognee have in common is that their memory capabilities are built on general-purpose storage, shielding underlying differences through an abstraction layer.
50%UnverifiedCognee, on its improve page, states its gates: guidance is eligible only when it has not been marked harmful and its confidence passes the distillation gate; proposed lessons are handed to the writer/rejecter to check their relationship with existing lessons and graph entities.
50%UnverifiedThe Cognee plugin, after each conversation turn completes, only writes the content to cheap session memory and does not build a graph; graph construction (chunking, LLM extraction of entities and relationships, embedding, and committing edges) is concentrated at the end of the session and performed in a separate process via cognee.improve()
50%UnverifiedThe Cognee local server runs in agent mode and automatically retires within 60 seconds after the last agent deregisters
50%UnverifiedCognee integration supports three connection modes: local-server (default), remote (set COGNEE_BASE_URL), and embedded (COGNEE_EMBEDDED=true).
50%UnverifiedCognee ingests content and builds a knowledge graph, enabling retrieval to traverse along relationship chains rather than returning isolated text chunks
50%UnverifiedThe Cognee plugin adopts a no-silent-fallback design: if COGNEE_BASE_URL is set but connection fails or the local server fails to start, initialization directly throws an error rather than switching to another mode.
50%UnverifiedSwitching models is fully decoupled from the memory layer, and Cognee treats different models such as Gemini and Opus the same at the underlying level.
50%UnverifiedTo install the Cognee plugin use the command pip install cognee-integration-hermes-agent, and upgrading requires adding the -U parameter.
50%UnverifiedCognee's local storage (SQLite, Kuzu/Ladybug, LanceDB) is single-writer, and multi-process access leads to database lock errors and eventually data corruption, so the official default adopts a local server as the single owner to serialize access.
50%UnverifiedCognee's local run mode requires an LLM API Key for extracting entities and relationships when building the graph; if Cognee Cloud is used, no LLM key is needed, but a tenant URL and API key are required
50%UnverifiedHermes discovers memory providers by scanning folders under $HERMES_HOME/plugins/ that contain plugin.yaml and a Python implementation; the folder name is the provider's identity
50%