Cognee vs OpenViking
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
OpenViking
Context database exposing agent memory, knowledge and skills as a filesystem
| What we compare | Cognee | OpenViking |
|---|---|---|
| Score parts, out of 100 | ||
| Adoption | 64, popular | 73, popular |
| Freshness | 100, active | 100, active |
| Maintenance | 82, healthy | 88, healthy |
| Easy to run | 67, easy | 67, easy |
| Agent-ready | 85, ready | 0, none |
| Facts from GitHub and the README | ||
| Stars | 32.1k | 39.8k |
| License | Apache-2.0 (permissive) | AGPL-3.0 (copyleft) |
| Last commit | Oct 2026 | Oct 2026 |
| Last release | Oct 2026 | Oct 2026 |
| Language | Python | Python |
| 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
OpenViking
OpenViking organises everything an agent knows as a viking:// virtual filesystem of resources, memories and skills, browsed with ls, tree, read and grep, with search scoped to a subtree. Each directory carries generated summaries so agents read full content only when needed. The server needs Python 3.10+ plus an embedding model and a VLM; plugins cover Claude Code, Codex, Cursor and OpenClaw.
Who it is for: Teams wanting transparent, file-like memory shared across coding agents
Strengths
- Memory is inspectable and editable as Markdown files under viking:// URIs
- LoCoMo accuracy 80 to 83% for OpenClaw, Hermes and Claude Code at far fewer tokens
- Python, Go and TypeScript SDKs plus HTTP API; multi-tenant accounts and ACLs
- Hosted Studio playground at openviking.ai/studio; Railway one-click deploy
Weaknesses
- AGPL-3.0 license
- Needs both an embedding model and a vision-language model from a provider
- Memory plugin installer is macOS/Linux only; Windows uses the beta desktop app
- Benchmarks were run with Volcengine Doubao models
- no GPU
- Docker + Compose
- Models: Volcengine, OpenAI, Codex OAuth, Kimi, GLM
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.