Pay for what you use

Final plans are being set now and will be published in full before the private beta opens. The shape is already decided:

Usage-based by default

GPU time is billed per second, application hosting by bandwidth and build minutes. Idle endpoints scale to zero and bill nothing.

No mandatory annual commitment to get a sane rate, and no charge for seats you are not using.

Usage and projected spend are visible in the dashboard as they accrue, with alerts you set yourself.

Waitlist members get pricing and beta access before it is announced publicly.