· Memory/ What it knows

Cognee

Cognee is a memory option for building AI agents. Builds a graph + vector memory layer over your agent's history and documents.

Open source · can be self-hosted · SDKs for Python · by Cognee · verified 2026-08-21

Official docsAdd to a stackAll memory options
· When to reach for it/ Fit
  • You are choosing a memory component — what it knows.
  • You need to run it on your own infrastructure, or your procurement process requires source access.
  • You want to start without a commercial conversation.
· What it takes to run/ Setup
Install
pip install cognee
· Cognee vs the alternatives/ 6 others
OptionWhat it doesLicenceSelf-host
CogneeBuilds a graph + vector memory layer over your agent's history and documents.Open sourceYes
Mem0Extracts durable facts from conversations and recalls them on the next turn.Open sourceYes
ZepTemporal knowledge graph memory — facts with validity windows, not just embeddings.Open sourceYes
HonchoTreats memory as a reasoning problem, not a retrieval one — you ask its Dialectic API a question about the user and get a synthesised answer.Open sourceYes
SupermemoryMemory graph plus automatic user profiles, with connectors that sync from Drive, Gmail, Notion and GitHub.ProprietaryNo
LangMemMemory primitives that plug straight into LangGraph checkpoints.Open sourceYes
Postgres (rolled by hand)A messages table and a summaries table. Boring, auditable, yours.Open sourceYes
· Stacks that use it/ 1
· For agents/ This page, machine-readable

Every page here answers to Accept: text/markdown and returns the same content at roughly a tenth the tokens. No separate site, no toggle — same URL.

curl -s -H "Accept: text/markdown" https://newagent.build/c/cognee