Short sign-ins, from places you allow.
A sign-in that never expires and works from anywhere is what actually goes wrong with AI agents. So the powerful ones here are the shortest lived, and you can say which countries and which networks any of them may be used from.
The strongest sign-ins last the least
Sign in from a terminal and what you get by default is read-only, good for ninety days. Ask for enough to run a project and it lasts thirty. Ask for full access and it lasts a week. A full sign-in that leaks costs you far more than a read-only one, so it is given far less time to be found. An AI agent gets less again, and how much less depends on how it was let in.
- Read-only: ninety days. Enough to run a project: thirty. Full access: seven.
- An AI agent a person let in with a code: seven days, expiring on its own whether or not anybody stops it.
- An AI agent your identity provider registers: five minutes, replaced as it works. Either way, stopping one acts on the AI agent rather than on any single sign-in.
- Signing in again retires the one before it for the same tool. The command line holds one sign-in per level of access, so a read-only sign-in is not retired by a wider one; a connected coding AI agent holds one per machine. Sign in from the command line on a second machine at the same level and the first machine is signed out.
- A machine sign-in can be set never to expire. We would rather you did not, and the default is ninety days.
Only from where you say
Your organization can say which countries and which networks any of its sign-ins may be used from. There are three ways in: a long-lived sign-in your organization mints, a key issued to a partner, and one issued to a named person or a named AI agent as they sign in. The place rules apply to all three, because a rule that held for two of them would read well and stop nothing.
- Creating a sign-in, or stopping one, asks for a real person signed in at that moment.
- A machine sign-in cannot do it however much else it is allowed, which closes the loop an AI agent would otherwise have: reaching the thing that governs it and issuing itself another.
A connection that runs away is slowed
One connection gets sixty tool calls a minute. Past that the next call is refused, told plainly to stop looping and batch the work, and the refusal is written to your record. An AI agent stuck in a loop is a bill and a mess, and it should hit a wall rather than a limit nobody notices until the invoice.
- Background work gets a sign-in of its own, and the previous one for that job is retired in the same step, so both are never live at once.
- The AI agent's name does not change with it, so the record still follows one AI agent across every run.
- There is no single button that swaps one sign-in for another: create the new one, move what uses it, stop the old one, in that order.
Ask us how long your AI agents' sign-ins should last.
Also in Studio
- StudioAgent runsGive an AI agent a job, then read back every step of it.Read the page
- StudioAgent securityA name, a limit, and somebody who can stop it.Read the page
- StudioAgent identityA person lets an AI agent in, and can stop it again.Read the page
- StudioWhat an AI agent may doIt never reaches further than its owner.Read the page
- StudioApproved modelsWhich models may run, who approved each one, and what it left behind.Read the page
- StudioScans and approved packagesApproved packages at the exact version, watched after that.Read the page
- StudioCommand lineThe terminal your engineers already use, on your rules.Read the page
- StudioMCP serverConnect any AI agent, and the rules come with it.Read the page
- StudioSDK and APIOne typed client for the whole platform, given to your team.Read the page