Cognee
Open source memory layer that turns documents and conversations into a knowledge store AI agents can query.
How the 8.2 was reached
- Quality of output8/10
31,222 GitHub stars, 3,133 forks, active recent pushes, Apache-2.0, and 5M+ SDK runs per month claimed. github.com
- Ease of use8/10
Installs with pip, runs locally with no API key, and has a docs site. github.com
- Pricing value9/10
Apache-2.0 open source, free local use, free cloud plan with 1M tokens, and $1 per 1M tokens after that. cognee.ai
- API and integration quality9/10
Python SDK, HTTP API and MCP, with integrations for Claude Code, Codex, Slack, GitHub and Linear. docs.cognee.ai
- Solves the problem it claims to solve7/10
Strong adoption (31k stars) supports the agent-memory claim; the team-memory features are shown only by vendor descriptions. github.com
8.2/10 is the mean of the 5 criteria that apply. Scored on the five-criteria rubric v1, 29 September 2026. How scoring works
Pricing
- 1M tokens included
- 1 workspace
- Unlimited users and API calls
- Claude Code, Codex and MCP integrations
- Unlimited workspaces
- Slack, Notion, Linear and Google Drive connectors
- Code indexing for repos
- In-app support
- Bring your own cloud
- Bi-temporal memory and conflict resolution
- Provenance on every answer
- Dedicated support and SLAs
Key Features
- Installs with pip; needs Python 3.10-3.14
- Runs locally with no API key using free small models
- Python SDK, HTTP API and MCP
- Integrations with Claude Code, Codex, OpenClaw and Hermes
- Shared team memory from Slack, GitHub and Linear data
Pros & Cons
Pros
- Apache-2.0 with about 31.2k GitHub stars
- Free local use and a free cloud plan with no card
- Simple token-based pricing
- Wide agent and tool integrations
Cons
- Team-memory features are shown only by vendor descriptions
- Data connectors sit on the paid Standard plan
- Enterprise pricing is not published
- Token-based billing needs watching as memory workloads grow
Cognee suits developers who want agents to keep context across sessions and would like to start free and local. Teams needing proven shared memory at scale should check the vendor claims against their own data. The 8.2 rating reflects clear pricing and strong integration.
Try Cognee →Added to scored.tools on
Competitors to Cognee
Other tools in the database category worth comparing.
Chroma
7.6/10Open-source vector database and search infrastructure for AI applications
Qdrant
8.2/10High-performance open-source vector database written in Rust for AI applications.
Weaviate
6.6/10Open-source vector database designed for AI applications and semantic search
Pinecone
7.4/10Vector database for building knowledgeable AI applications with fast similarity search