Security and your data
What is kept, who can see it, and how to end it.
- Your agents
- What each is for, what it may do, every version, and the conversation that produced it.
- Your runs
- What you asked for, every step taken, what went in and came back, and what it cost.
- Connected accounts
- An encrypted token where a service gave one. Never a password. A site login is cookies in the agent's own browser, and all that is noted is which browser.
- Files
- Anything an agent produced for you, until you delete it.
- Sign-ins
- A browser, a time and the address it came from, per device.
Everything is scoped to a workspace, and every page and every route checks your role on the server before reading anything. There are three roles. An Owner can do everything including billing, members and deletion. An Editor can build, run and connect, but not change billing or members. A Viewer can read and start nothing.
One deliberate exception runs the other way: an Owner cannot see anybody else's signed-in devices. Reading a colleague's IP addresses is surveillance, and being an Owner is not a reason for it.
Role changes, connections, device sign-outs and exports are recorded with who did them. The log is append-only in the database rather than by convention — the table refuses updates and deletes, so not even an Owner can edit it. Exporting it is itself recorded, because somebody taking a copy of who did what is exactly what the next reader wants to know about.
Sign out everywhere ends every other session immediately and leaves the browser you are using alone.
Not live yet. Deleting your data is not built. Removing an account means removing runs, steps, files, tokens and the rows the sign-in library owns, in an order that strands no reference — get that wrong and somebody is left half deleted. The button records a dated request and says so. Nothing is removed by it.