Priced for people. Ready for agents.

You pay per user, monthly or annually, with every feature included. Each person is on Business or Business Premium, and a workspace can mix both. Users you add are prorated onto your next invoice; if you reduce the number of users, you pay until the end of the period. Pending invitations are not billed.

Yes, with Enterprise. The same full stack runs on your own servers or in your own datacenter with Docker Compose, and model usage goes through your own provider accounts, billed by them directly. Enterprise also covers volume pricing and an annual contract paid on invoice.

A full workspace with every feature, for up to 5 people, with no card needed. When it ends, everything stays readable and exportable, and chat and workflows start again as soon as an admin subscribes.

A lighter model answers until the limit resets, and the person can ask an admin for more. The admin can raise their limit, move them to Business Premium, which includes 5 times the usage of Business, or turn on extra usage. If the workspace's included usage for the month runs out, a lighter model keeps answering, with limited capacity, until the 1st of the next month.

It's off until an admin turns it on. People then keep their chosen model beyond what's included, up to the monthly caps you set: for the workspace, for each person, and for workflows, agents and API keys. Admins are told at 80% and 100% of the cap, and extra usage appears as its own line on your next invoice.

Models from OpenAI, Anthropic, Google and Mistral, in one menu. Admins choose which ones the workspace offers, and can connect their own model-provider keys.

No. Both are included for every user. Workflows include 1,000 runs a month per workspace, and the API allows 600 requests a minute per workspace. Talk to us if you need more.

Prices exclude VAT. Add your company's VAT number at checkout. Invoices, your payment method and cancelling the subscription are in Settings › Billing › Invoices and payment.