One shared memory, with roles.
An organization workspace with owner, admin, and member roles over a single shared memory. Issue per-org API keys for connected agents — everyone works against the same graph.
Owner, admin, and member — one shared graph, with per-org API keys. Not a SQLite file stuck on one laptop.
Owner, admin, and member roles plus per-org API keys all converge on one shared memory.
One canvas, edited together
You and your team's agents work the same shared memory — like cursors on one canvas. What anyone writes, the next agent can read.
One shared memory
roadmap
editingYou&Claude
An illustration of the idea: one shared graph, many editors. Real agents read and write over MCP — the same single source of memory, not literal live cursors.
Roles over one shared memory
An organization workspace gives your team owner, admin, and member roles over a single shared memory. Everyone on the team works against the same graph — not a SQLite file stuck on one person's laptop — so what one agent learns is available to the rest.
- Owner, admin, member — roles over one shared memory
- One graph for the team — not a SQLite file stuck on a laptop
- Shared by default — what one agent learns, the team can reach
Keys for your agents, stored hashed
Connected agents authenticate with per-org API keys. A key is shown once when you create it and stored hashed after that, so the secret never sits in plain text. Every connected agent works against the same shared memory through its key.
- Per-org API keys — shown once, stored hashed
- One memory for every agent — keys scope access to your org's graph
More capabilities
Five capabilities, one shared memory. Here are the other four.
One memory for the whole team.
Start free and bring your team onto one shared, governed memory for your agents.
No credit card required · Free plan available · Bring your own OpenAI key