Help Seats, models and keys

"Runs on": Our Keys, BYOK, Custom Endpoint

Each seat chooses how it is powered, separately from the account's default.

Every seat has a Runs on setting with three choices. It is per seat, so one chair can run on your own key while the next runs on ours.

Our Keys

The seat runs on our provider keys and draws down your credit, billed at what the provider charges us plus 10% for serving the call. Nothing to set up: an account with credit is enough. Your prompts, and any file contents you attach, pass through our servers on their way to the provider. This is the default for a new account, which starts with $2 of credit.

BYOK

The seat runs on an API key you add — Anthropic, OpenAI, Google or xAI. The provider bills you directly and your credit is untouched. On the web and the phone the key is encrypted at rest and used only on our servers to run your discussions. On the Mac the key stays on the Mac, encrypted there, and the Mac calls the provider itself.

Custom Endpoint

The seat talks to an OpenAI-compatible endpoint you name, with that endpoint's own key. Our servers are not involved and credit does not apply. See "Putting any model in a seat".

The account's default

Settings → Plan & Billing has the account-wide choice between Our Keys and your own. A seat that has made no choice of its own follows it. A seat set to BYOK with no key saved follows the account too — a choice with nothing behind it is no choice.

CLI seats

On the Mac a seat can also be a CLI — Claude Code, Codex or Gemini CLI — running on the subscription you already pay for. That is a fourth way to power a seat and bills nothing per call; see "CLI seats". On the free plan a seat on your own key, a custom endpoint, your company's cloud or a CLI works for one 14-day trial, then needs Pro; Our Keys and Free Key have no such limit.

Still stuck?

Open Get help in the app and send us a message, or see the support page for how to reach a person.