Home / Directory / Agents & orchestration / AI agent memory / Cognee
Cognee
Open-source memory engine from Berlin that turns documents, chats, and code into a knowledge graph agents can search, with a token-priced cloud and in-your-cloud deployments.
Engineering teams that want open-source graph memory they can self-host, and companies that want help building domain-specific memory inside their own cloud.
Buyers who want a fully managed enterprise service with a public price, or teams that only need simple chat history.
Verdict
Cognee is worth using for teams that want open-source graph memory they can run themselves, with a managed cloud option when they are ready. It suits companies building domain-specific agents with their own rules and data.
Score Breakdown
How Cognee scores in the categories that matter to its buyers.
Pricing
Cognee publishes token-based cloud prices, and its open-source engine is free to run on your own infrastructure.
| Plan | Price | What stands out |
|---|---|---|
| Free | $0 | 1M tokens, 1 workspace, unlimited users and API calls |
| Standard | $1.00 / 1M tokens + $5 per added workspace a month | Slack, Notion, Linear, and Google Drive connectors, code indexing |
| Enterprise | Custom BYOC engagement | 6, 12, or 24 months in your cloud; discounted startup package for pre-Series B companies |
The Field at a Glance
How Cognee compares with other AI agent memory vendors we reviewed, by Overall Score and relative typical engagement cost.
Cognee scores 7.7 in AI agent memory, under Mem0 (8.0), Zep (7.9), Letta (7.8), and Supermemory (7.8). Relative cost lands lower for this subcategory.
Use-case matrix
| Use case | Fit | Notes |
|---|---|---|
| Graph plus vector memory | Strong | Relational, vector, and graph storage sit in one engine. |
| Self-hosting | Strong | The open-source engine runs locally or on your stack. |
| Low cloud price | Strong | Standard costs $1 per million tokens processed. |
| Agent integrations | Strong | Works with Claude Code, Cursor, LangGraph, and MCP clients. |
| Managed enterprise plan | Mixed | Enterprise is a scoped engagement with no public price. |
| Simple chat memory | Mixed | More setup than a basic history store. |
Who it’s for
Good fit
- Teams wanting open-source memory
- Vertical AI companies with domain rules
- Research and education groups
Poor fit
- Buyers wanting a priced managed plan
- Simple chat history needs
- Teams without engineering time
Review Excerpts
Below are excerpts from public reviews. Our team scoured public reviews, forums, and chat rooms to get a balanced view of customers' experience with this company. Paid reviews and pay-for-play sites such as Clutch were excluded.
“We tried understanding the proximity and relationships of our users before, but SQLs got out of hand and embeddings were too sparse. With cognee, we managed to get a POC done in 2 days on 40,000 students from Bremen.”
“Cognee's design allowed our team to implement AI memory capabilities on-prem without worrying about graph or vector database configuration. The accuracy of our information retrieval has significantly increased.”
Cognee's Enterprise plan is a fixed-scope engagement of 6 to 24 months in the customer's cloud with no public price, and bi-temporal memory and answer provenance sit on that tier.
Company Facts
Key Leadership
Vasilije Markovic
Founder and CEO
Methodology
This page is an independent evaluation of Cognee for buyers comparing options in AI agent memory. AI Industry Reviews accepts no sponsorships, advertising, or pay-for-placement fees. Cognee did not pay for this review.
What we scored
The headline number is an Overall Score on a 0-10 scale. Eight criteria sit under it in two groups.
- Buyer outcomes (for AI agent memory)
- Knowledge graph memory
- Open-source and self-hosting
- Data connectors
- Managed cloud maturity
- Company & commercial
- Innovation & product leadership
- Project management & communication
- Pricing
- Contract fairness
Pricing measures whether the price looks fair for the value delivered, including packaging and renewal friction that show up in real buying cycles.
Score Composition
| Input | Weight | What it covers |
|---|---|---|
| Reviews | 40% | A proprietary read of what practitioners say about likes, complaints, and day-to-day use, including public review sites, forums, and private chat rooms. Paid reviews and pay-for-play sites such as Clutch are out of scope. |
| Product | 35% | Hands-on look at screens and workflows. |
| Pricing | 15% | Whether the price looks fair for what you get. |
| Docs & training | 10% | Docs, tutorials, and training material. |
How we balanced the evidence
The Overall Score is the simple average of the eight criteria. Recommendation language follows that score and the fit pattern described above.
Scope
Cognee is graded here as AI agent memory. Criteria scores can move as more review volume and product checks are added.