Cognee vs agentmemory
Two of the top memory, side by side: score, setup, license, activity and what each review found.
Cognee
Memory engine that turns documents and code into a knowledge graph
agentmemory
Persistent memory server for coding agents, exposed over MCP and REST
| What we compare | Cognee | agentmemory |
|---|---|---|
| Score parts, out of 100 | ||
| Adoption | 64, popular | 48, known |
| Freshness | 100, active | 100, active |
| Maintenance | 82, healthy | 81, healthy |
| Easy to run | 67, easy | 83, very easy |
| Agent-ready | 85, ready | 30, minimal |
| Facts from GitHub and the README | ||
| Stars | 32.1k | 29.3k |
| License | Apache-2.0 (permissive) | Apache-2.0 (permissive) |
| Last commit | Oct 2026 | Oct 2026 |
| Last release | Oct 2026 | Oct 2026 |
| Language | Python | TypeScript |
| Docker | Yes | Yes |
| GPU | Not needed | Not needed |
| arm64 or Apple Silicon | Mentioned | Mentioned |
Cognee
Cognee builds persistent agent memory by extracting entities, relationships and chunks from text, code and sessions into a graph and vector index with hybrid recall. Without an LLM key it uses local GLiNER extraction and embeddings; adding a key enables generated answers via OpenAI, Ollama or other providers. It runs as a library, CLI, REST API (port 8000), UI (3000) and MCP server (8001).
Who it is for: Developers building graph-backed memory for agents and company knowledge
Strengths
- Keyless mode: local GLiNER extraction and embeddings, no cloud LLM required
- Claude Code and Codex plugins, OpenClaw plugin, MCP server, Python, TypeScript and Rust SDKs
- Imports memory from Mem0, Letta, Zep or Graphiti via the COGX format
- Apache-2.0; research paper and BEAM evaluation published
Weaknesses
- Single-Postgres graph store is a demo; production version is a licensed product
- API defaults to multi-tenant mode; local use needs ENABLE_BACKEND_ACCESS_CONTROL=false
- Bundled GLiNER extractor is a demo; higher-accuracy version requires contacting the vendor
- UI launcher needs Node.js/npm and Docker for its MCP service
- no GPU
- Docker + Compose
- Compose runs Neo4j, PostgreSQL, Redis
- Models: local GLiNER + embeddings (keyless), OpenAI, Ollama, other providers per docs
- port 8000
agentmemory
agentmemory captures what a coding agent does across sessions, stores it as searchable memory, and injects relevant context at the start of the next session. It runs as a local Node.js server on the pinned iii engine and connects to agents through hooks, MCP, or REST, with 20 adapters listed. Keyless mode uses BM25 search; vector embeddings need a provider or the local Xenova/all-MiniLM-L6-v2 model.
Who it is for: Developers using coding agents who want memory kept across sessions
Strengths
- No external database; state lives in a local iii engine data directory
- Works with 20 listed agents through hooks, MCP, or REST
- Keyless BM25 mode works without any API key
- Local embeddings via EMBEDDING_PROVIDER=local after a one-time model download
Weaknesses
- Keyless mode has no vector search, so semantic queries can return nothing
- Pinned to iii-engine v0.22.1; refuses to attach to other engine versions
- Native Windows needs manual iii.exe install; WSL2 or Docker recommended
- Uses four local ports (3111, 3112, 3113, 49134)
- no GPU
- Docker + Compose
- Needs Node.js 20+, iii-engine v0.22.1
- Models: Xenova/all-MiniLM-L6-v2 (local embeddings)
- port 3113
More in Memory
Building with these?
We help teams pick an open-source AI stack and run it in production. Client work never changes a rank.