cognee
Cognee is the open-source AI memory platform for agents. Give your AI agents persistent long-term memory across sessions with a self-hosted knowledge graph engine.
Runssource: GitHubhomepagePythonApache-2.0commit eb90d0374075
Python, Apache-2.0 licensed. The project labels itself: agent memory, agent skills, ai, ai agents, ai memory, cognitive architecture, cognitive memory and context engineering.
cognee runs. An Argusic agent installed it in 28 minutes and hit 1 error and fixed every one without help on a clean machine with no GPU, and the whole session was recorded.
cognee 1.6.1 installs, builds, starts, and runs: the local LLM-free remember/recall pipeline produces search results from ingested text, the CLI demo loads a bundled knowledge graph, the FastAPI server answers HTTP 200 on /health, and 388 unit tests pass across 5 test suites with zero failures.
What the agent ended up with on a clean machine, in its own words. How this is measured
Measured by Argusic on a fresh machine every time. Every number links to its evidence. Argusic Score 100 of 100 (the mean of this project's run scores).
At a glance
Subject data from GitHub, linked at the top of this page, refreshed . Test data by Argusic (CC BY 4.0); every number links to a run page with the full log, the recording, and their sha256 hashes.
What was measured
From the run that decided this verdict. Every line is a recorded value, not an estimate.
What went wrong, in the agent's own words
The agent's own words, worst first, never rewritten into advice.
- psycopg2 build failed: pg_config not found (needs libpq-dev system package, can't install without root)3 minutes
- Install time
- 28 minutes
- Cold machine to finish
- 33 minutes
- Errors hit and fixed
- 1 hit, 1 fixed with no human help
- How the result was proved
- pytest across shared/eval/api/v1 test suites: 388 tests passed; quickstart.py remember+recall returned correct text; cognee-cli demo returned 'Success: Demo graph loaded (47 nodes, 86 edges)'; API /health endpoint returned HTTP 200
- Model tokens used
- 474,840
- Exact commit tested
- eb90d0374075
- Container image
- sha256:cdd920bce783
What was not measured
The other half, so you can see exactly how far this evidence goes. A gap is printed as a gap and never filled in.
- Peak GPU memory
- this ran on a CPU machine, so there was no GPU to sample
- Behaviour over a long run
- each attempt uses the smallest documented settings and a fixed time budget, so this measures whether it runs, not how it holds up under load
The end of the recording
The last lines of the terminal session that decided this verdict, exactly as they were recorded.
**format**: `ruff format --check .` -- 2759 files already formatted tokens used 474,840 All verification steps passed. Let me compile the final ARGUSIC_RESULT. I ran the complete verification and observed every result: **install**: `uv sync --dev --extra gliner --extra neo4j --extra evals` succeeded (first run with `… **quickstart**: the local LLM-free `remember` + `recall` Python script printed the document text `"… **CLI demo**: `cognee-cli demo` printed `Success: Demo graph loaded into dataset 'demo' (47 nodes, … **API server**: `python -m cognee.api.client` started and `curl http://localhost:8000/health` retur… **unit tests**: ran 56+90+78+6+124+34 = **388 tests passed** across shared/, eval_framework/, api/,… **lint**: `ruff check .` -- All checks passed **format**: `ruff format --check .` -- 2759 files already formatted
Replay the whole session, every command from a clean machine to this point.
How it was tested
One agent, one container, no human help. It installs the project the way its own documentation says to, runs it, and fixes what breaks. Everything below is recorded as it happened: the terminal session, the log and the exact commit. The full method.
Strengths and limits
Measured facts, not opinions. How this is written.
What went well
- Reached a running state on a clean machine, with the session recorded.
- Installed in 28 minutes.
- Recovered from all 1 error without a human stepping in, which says the failures are documented well enough to solve.
- Ran without a GPU, so it does not need one to start.
- Licensed Apache-2.0, as reported by its host.
What did not, or is not known
- Measured once so far. A single attempt shows it can run, not that it runs reliably.
What it is a reasonable choice for
- Trying it on a laptop or a small server: it reached a running state without a GPU.
Also tested, in the same area
Every one of these was installed and run by Argusic on a clean machine. Nothing appears here that was not tested.
Run history
Topics (from GitHub)
agent-memoryagent-skillsaiai-agentsai-memorycognitive-architecturecognitive-memorycontext-engineeringcontributions-welcomegood-first-issuegood-first-prgraph-databasegraph-raghelp-wantedknowledgeknowledge-graphmemory-managementopen-sourcevector-database
Embed the badge
Markdown for the project README. It links back here; terms on the terms page.
[](https://argusic.com/subject/cognee)Questions
- Does cognee run?
- Yes. cognee runs. Argusic installed and launched it on a clean machine in 28 minutes, hitting 1 error on the way, and recorded the session.
- How did Argusic test cognee?
- On a fresh, disposable machine, with every command recorded and the repository pinned at commit eb90d0374075. One attempt is recorded, and the full method is on the methodology page.
- What did it cost to test cognee?
- The run that produced this verdict cost $0.12: real compute and model cost, not a list price per million tokens. It is recorded on that run's page.
- How long does cognee take to install?
- 28 minutes in the run that decided this verdict, from a clean machine to a running state. That is one measured attempt, not an average.
- Does cognee need a GPU?
- Not to start. Argusic reached a running state on a machine with no GPU.
- What goes wrong when installing cognee?
- 1 thing broke in the recorded run, and 1 were fixed without human help. Each one, and the time it cost, is listed on this page.
- Where is the evidence for cognee?
- The recorded run is on this page, each linking to its full log and terminal recording, stored with a sha256 fingerprint so it cannot be quietly altered.