Cognee
"Extract–Cognify–Load" pipeline that turns raw input into a typed, queryable knowledge graph for agent recall.
Active · Verified 2026-10-22 · fresh as of 2026-10-22 · 72 cited facts, 6 with a third-party source · 1 recorded change · Suggest a correction
At a glance
- Type
- Knowledge graph + ECL pipeline vendor fix
- Tier
- T1
- Section
- Dedicated memory layers
- Latest release
- v1.6.5 (2026-10-16) third-party fix
- License
- Apache-2.0 repo fix
- Pricing
- Free $0 (1 workspace, 1M tokens); Standard $1 per 1M tokens + $5/workspace/mo; Enterprise BYOC custom vendor fix
- Funding
- $10M total Seed · 2026-02 third-party fix
- Freshness
- 2026-10-22 fix
- Layer
- agent-infra repo fix
- SDK languages
- python, rest vendor fix
- Protocols
- mcp-server vendor fix
- Developer activity
- OpenRank 25.91 (12-mo avg 26.62) · 125 participants — 2026-09 third-party fix
Memory taxonomy
- storage
- graph
- retrieval
- graph-traversal
- persistence
- long-term
- update
- extraction
- unit
- fact
- governance
- inspectable
- conflict
- llm-arbitrate
When to use
Optimised for: typed knowledge graph extraction (ECL pipeline) over RAG
Anti-fit: no anti-fit explicitly stated
Pros & cons
Pros
Pipeline-as-code ECL (extract/cognify/load) makes the memory build path inspectable and replayable; fully OSS.
Cons
Smaller community than Mem0/Letta; more end-to-end engineering required to deploy.
Claims & capabilities
€7.5M seed Feb 2026. Native integrations with Claude Agent SDK and Google ADK.
Technical surface
- API surface
- REST, SDK: Python
- Backend storage
- hybrid (vector + graph)
- Deployment
- Both (OSS self-hosted; Cognee Cloud managed; Enterprise BYOC)
- Embedding model
- multiple supported
- Multi-tenancy
- Logical namespace per project/dataset; self-hosted OSS deployment
- MCP
- native (first-party) — cognee-mcp
- A2A
- not documented publicly
- OpenTelemetry
- not documented publicly
Compare Cognee with…
Alternatives to Cognee → live options ranked by shared design, licence, hosting and MCP support
Similar systems
Other dedicated memory layers in the catalog, ranked by inbound references.
- Mem0 T1
Universal memory layer for AI agents. Three concurrent stores (vector + graph + KV); LLM-extracted facts; concurrent retrieval via ThreadPoolExecutor.
- Zep & Graphiti T1
Bi-temporal knowledge graph (event time + ingestion time). Strong on chronological reasoning and contradiction tracking. Graphiti is the open-source core.
- Hindsight (Vectorize) T1
Standalone memory service from Vectorize. Open source. Biomimetic four-network design (World, Bank, Observation, Opinion). Ships an MCP memory server.
- Letta / MemGPT T1
Originated as MemGPT (UC Berkeley), where the model pages its own memory between context, recall and archival tiers. The product now centres on Letta Code, a stateful agent harness whose agents rewrite their own memory, skills and prompts over time.
- Memori T2
Structured, persistent memory layer for AI agents. Open source.
- Memvid T2
Single-file memory layer (one .mv2 file). No DB, no server. Append-only sequence of immutable Smart Frames with timestamps + checksums. Native Rust core (rewritten from Python).
Sources (71 cited facts, 6 third-party)
Each fact links to the page it came from; the badge says who published it (how to read sources).
- What it is Knowledge graph + ECL pipeline vendor
- What it does & what's distinct "Extract–Cognify–Load" pipeline that turns raw input into a typed, queryable knowledge g… vendor
- Claims / benefits €7.5M seed Feb 2026. Native integrations with Claude Agent SDK and Google ADK. vendor
- Modalities text vendor
- Created 2023-08 vendor
- Latest release v1.6.5 (2026-10-16) third-party
- License Apache-2.0 repo
- GitHub topoteretes/cognee — ~31.2k stars, last commit 2026-10 repo
- Mindshare 2 inbound 11 HN repo
- Funding $10M total Seed · 2026-02 third-party
- Customers / scale 11-50 ppl 70+ companies in live environments (unnamed) third-party
- Pricing Free $0 (1 workspace, 1M tokens); Standard $1 per 1M tokens + $5/workspace/mo; Enterpris… vendor
- Deployment Both (OSS self-hosted; Cognee Cloud managed; Enterprise BYOC) vendor
- HQ DE vendor
- Founders / pedigree EU startup (Berlin); co-founders Vasilije Markovic (data engineering lead at Zalando / T… third-party
- Memory volume / scale no scale claim vendor
- API surface REST, SDK: Python vendor
- Backend storage hybrid (vector + graph) vendor
- Multi-tenancy Logical namespace per project/dataset; self-hosted OSS deployment vendor
- Embedding model multiple supported vendor
- Consistency eventual vendor
- Versioning snapshots vendor
- Tombstoning soft-delete only vendor
- Schema evolution schema with migration vendor
- Namespace primitives per-project vendor
- Contradiction handling LLM resolves and overwrites vendor
- Forgetting policy user-controlled forget vendor
- MCP support native (first-party) — cognee-mcp repo
- Import / export JSON; Neo4j / Postgres backends pluggable vendor
- Integration count 1 repo
- Orchestration both (single-agent default; multi-agent via dataset/user scoping) vendor
- Programmatic control full CRUD + SDK (Python) vendor
- Pricing specifics per-API-call + tier flat-rate: Free; Pro €40/mo; Enterprise custom vendor
- Validated verticals internal IT, knowledge management, research vendor
- Time to running <1hr (pip install + Docker for graph store) vendor
- Anti-fit no anti-fit explicitly stated vendor
- Optimised for typed knowledge graph extraction (ECL pipeline) over RAG vendor
- Adjacent infrastructure BYO LLM; bundles Kuzu/Neo4j for graph + LanceDB for vectors vendor
- Pros Pipeline-as-code ECL (extract/cognify/load) makes the memory build path inspectable and … vendor
- Cons Smaller community than Mem0/Letta; more end-to-end engineering required to deploy. vendor
- Project & sources GitHub topoteretes/cognee how Cognee builds memory Memify pipeline Claude SDK integration vendor
- obs-langfuse yes third-party
- cost-token-budget no vendor
- cost-prompt-caching no vendor
- cost-semantic-caching no vendor
- cost-batching no vendor
- cost-model-routing yes vendor
- cost-streaming-only no vendor
- cost-observability-cost-attribution no vendor
- eval-langsmith-evals no vendor
- eval-braintrust no vendor
- eval-weights-and-biases-agent no vendor
- eval-helicone-evals no vendor
- eval-custom-test-harness yes repo
- eval-human-loop no vendor
- eval-production-traffic-replay no vendor
- commit-trajectory [{"ym":"2026-03","cum":6462},{"ym":"2026-04","cum":7081},{"ym":"2026-05","cum":7170}] repo
- download-trajectory [{"ym":"2025-11","monthly":10773,"cum":10773},{"ym":"2025-12","monthly":23709,"cum":3448… registry
- cost-input-usd-per-mtok 0 vendor
- cost-output-usd-per-mtok 0 vendor
- cost-tier free vendor
- cost-pricing-model free-tier vendor
- cost-last-verified 2026-06-29 vendor
- architecture-deployment-locus hybrid vendor
- licensing-bucket open-core vendor
- positioning-platform-vs-product platform vendor
- taxonomy-three-layer-rollup agent-infra repo
- developer-activity-opendigger {"activity":332.49,"as_of":"2026-09","openrank_12mo":26.62,"openrank_current":25.91,"par… third-party
- sdk-languages python,rest vendor
- protocol-support mcp-server vendor
- maturity beta registry
What changed
Changes to this record's lifecycle, licence, release, pricing and hosting facts, from the catalog's git history. Subscribe (Atom)
- 2026-10-22
- Latest release “v1.0.8 2026-05-06” → “v1.6.5 (2026-10-16)” source
- Pricing “Free + paid” → “Free $0 (1 workspace, 1M tokens); Standard $1 per 1M tokens + $5/workspace/mo; Enterprise BYOC custom” source
- Deployment “Both” → “Both (OSS self-hosted; Cognee Cloud managed; Enterprise BYOC)” source
Related systems
References (2)
- LanceDB depends on at runtime — adjacent-infrastructure cell: BYO LLM; bundles Kuzu/Neo4j for graph + LanceDB for vectors
- Neo4j depends on at runtime — adjacent-infrastructure cell: BYO LLM; bundles Kuzu/Neo4j for graph + LanceDB for vectors
Referenced by (8)
- Amazon Neptune Analytics integrates with — Cognee integration for agentic RAG
- Cognee MCP builds on — Cognee MCP — Bundled in Cognee v0.3.5+
- Haystack (framework) integrates with — Haystack integrations directory has a Cognee integration page ("Add persistent, knowledge-graph-backed memory to your Haystack agents and pipelines with Cognee"). (verified 2026-10-29)
- LangChain (framework) integrates with — LangChain's all-providers integrations page lists Cognee: "Memory layer for AI applications and agents". (verified 2026-10-29)
- Langfuse integrates with — Langfuse's integrations directory has a Cognee integration page. (verified 2026-10-29)
- LlamaIndex (framework) integrates with — LlamaIndex has an official graph RAG integration package llama-index-graph-rag-cognee in its repo. (verified 2026-10-29)
- Milvus integrates with — Milvus's integrations overview lists a Milvus + Cognee integration tutorial. (verified 2026-10-29)
- Qdrant integrates with — Qdrant's documentation has a Cognee integration page. (verified 2026-10-29)