Pricing
You pay for the work the AI agent performs and for the infrastructure your application consumes — itemized separately, so a month's total is always explainable.
Pay per workperformed
Measured by the work the agent actually does: analyzing a repository, building and verifying a release, diagnosing an incident, performing a repair or redeploy. Idle applications generate no agent charges.
Pay per hourprovisioned
Managed components are billed for the time they exist, at a flat hourly rate per size. Stop or delete a component and billing stops with it.
| 1 × app server (starter) | $12.00 |
| 1 × managed PostgreSQL (starter) | $15.00 |
| 1 × managed Redis (starter) | $8.00 |
| Agent usage — ~8 deployments, 2 incident repairs | ≈ $20.00 |
| Estimated monthly total | ≈ $55.00 |
Figures are illustrative examples for a small web application with one service, one database and one cache. Actual totals depend on component sizes, uptime, and the amount of agent work performed. Final early-access rates are confirmed during onboarding.
Bundling agent work into an infrastructure price hides what you're paying for. Keeping the meters separate means a quiet month with a stable application costs roughly the infrastructure alone, while a busy month of deployments shows up clearly on the agent meter.