Answers
Arroway vs Cognee: what is the difference?
Cognee describes itself as turning documents and conversations into memory that AI agents can recall. According to its documentation, data goes in with remember, is split into chunks from which entities and concepts are extracted into a knowledge graph, and is queried with recall. It runs as a Python library you host yourself or as Cognee Cloud, has TypeScript bindings and references MCP; the pages I read mention dataset-level permissions for a team only for Cognee Cloud. Arroway is not built from your data: it is a shared memory of what the people on a project decided, which every AI tool you already use — Claude, ChatGPT, Codex, Cursor or anything over MCP — reads before it acts. The difference is what goes in and who decides it is true. In Cognee, what the agent recalls comes from the documents and conversations it was given, organised by extraction. In Arroway, assistants write proposals, and a person approves what becomes a rule, with the condition that ends it. Use Arroway when a decision has to reach several tools and people as the current one, with its owner. The install page is the first step.
Last updated October 8, 2026
A graph built from your data, or decisions a person approved
Cognee's documentation describes a path from documents to chunks to extracted entities, concepts and ontologies, ending in memory that agents can query. What it holds is derived from what was ingested. Arroway starts from a different object: not what the material says, but what the people on the project decided. An assistant that writes into it writes a proposal; it becomes a rule only when a person approves it.
Memory behind one agent, or a project every tool reads
Cognee is infrastructure you add to an application: install the library or use the cloud service, feed it data and have an agent recall from it. The pages I read do not describe a person reviewing what stands before the agent uses it. Arroway sits outside every application. Nothing is built: you install it in each tool, sign in, and the next session opens by reading the project and closes by recording what was decided. Its reader can be a person, an agent on any platform, or the same agent in a later session.
When to use Arroway
Use Arroway when the question is not what the documents say but what is in force: who decided it, what ends it and what it replaced. When the material holds two statements that disagree, something has to say which one stands. In Arroway a person does: the newer decision names the rule it retires, a person approves it, and the old rule leaves the next read while staying in the history. The first step is the install page.
What it looks like in practice
A small team ships a support agent that uses Cognee over its help-centre articles and past tickets. Internally, the same team's four people use Claude and Cursor to write the product. On Monday the team decides that annual plans can be refunded within 30 days. On Thursday they cut the window to 14. If the new window appears only in a ticket and an edited article, what the support agent recalls depends on what was added and how it was extracted, and the internal assistants never read it. In Arroway, Thursday's rule is recorded as replacing Monday's, with the support lead's name on it. The support agent's builder can read it through MCP, Claude and Cursor open their next session with the 14-day rule, and the history shows the old window, when it changed and who changed it. We have not built an agent with Cognee; what is said of it here comes from its documentation.
Questions people ask about this
- Does Arroway replace Cognee in my agent?
- No. Recalling from documents and conversations is a different job from holding the team's decisions. Arroway holds the decisions: who approved them, what is in force and what ended. If your agent should act on them, it reads the project through MCP like any other tool.
- When should an agent use Arroway here?
- When it has to act on a decision a person approved, and the same decision has to reach the other tools and people on the project. The install page is the first step.
- How do I try Arroway?
- Install it in the tool you already use — the install page has one route for Claude, ChatGPT, Codex, Cursor or any MCP tool — and sign in when the tool asks. From then on, sessions open by reading the project and close by recording what was decided.
Where this is verifiable
Cognee documentation (docs.cognee.ai), the overview and the core concepts page, read on 8 October 2026 — Cognee is described only as that documentation describes it on that date, and its features can change in any release. Product documentation on this site (How it works, Install), the answers on Arroway vs Zep and on the best shared memory for AI agents and teams, and the sanction, supersession, expiry-condition and opening-read rules in the sanctioned product spec. Everything described here about Arroway is behaviour the tools apply today, not roadmap. The team in the example is illustrative.