Skip to main content
You pay for what your app uses. There is no per-seat charge on usage and no charge for idle apps that have scaled to zero.

​
What gets metered

Managed Postgres, Storage (stored data and operations), bandwidth and egress, and builds are included at no charge today. They do not appear on your usage ledger. Exact unit rates are published at getfloo.com/pricing, effective-dated so past invoices stay reproducible. One floo credit is $1 of metered usage.

​
Plans

Every plan can run every primitive: web services, workers, cron, managed Postgres, Redis, and Storage. Plans gate how much you can spend and who can administer it, never what you can build.

​
Spend caps

Set a monthly cap, in dollars, so an agent cannot run up a bill. The cap is org-wide, not per app. 0 turns budget protection off.
A cap comes with a policy that decides what happens at 100%: The CLI sets only the amount. floo billing spend-cap set keeps the policy the org already has, or selects freeze_new_spend when budget protection was off. Pick a different policy on the dashboard Billing page. Under freeze_new_spend and hard_stop, a blocked deploy fails with the typed error SPEND_CAP_EXCEEDED. Warning emails go out at 80% under every policy. See Billing and spend caps for the full workflow.

​
What this costs in practice

Under the floo-credits-2026-09 rate policy, effective 2026-09-01, CPU is 0.0468pervCPU−hourandmemoryis0.0468 per vCPU-hour and memory is 0.0216 per GiB-hour. One warm instance for 730 hours:
  • Default HTTP shape, 1 vCPU / 512 MiB: 730 × 0.0468+730×0.5×0.0468 + 730 × 0.5 × 0.0216 = 34.16+34.16 + 7.88, about $42/month.
  • The same service attached to managed Postgres runs a 0.5 vCPU / 256 MiB database proxy next to it, 1.5 vCPU / 768 MiB in total: 730 × 1.5 × 0.0468+730×0.75×0.0468 + 730 × 0.75 × 0.0216 = 51.25+51.25 + 11.83, about $63/month.
Both figures assume the instance stays warm; a service that scales to zero is billed only for what it serves. See Scaling and availability to choose a runtime posture.