Operate
Usage and billing
See what your organisation used and what it cost, by day, by resource, by project and by unit, and how each amount is worked out.
On this page
Where to look#
Open Usage at the bottom of the sidebar, next to Team. It covers the whole organisation: every project, environment, unit, volume and bucket. A project's own Usage page shows the same spend for that project alone, next to its CPU and memory graphs.
kuiper usage
kuiper usage --month 2026-09
kuiper usage --from 2026-09-01 --to 2026-09-15
kuiper usage -p proximakuiper usage prints this month's spend so far, where the month is heading, each resource with its price, and a line per project. --month and --from/--to choose another period (both days are included); --json prints everything the API returned. With -p it shows one project's reserved and used CPU and memory and its spend.
What the page shows#
| Part | What it tells you |
|---|---|
| Spend | What the period cost so far, in US dollars. |
| Month to date and Projected month end | This month so far, and where it ends if the last day's hourly rate carries on. |
| Spend per day | One bar a day, split by resource. Hover a bar for the amounts. |
| By resource | How much of each resource was billed, how much was actually used, the price, the spend and its share of the total. |
| By project | Each project with its CPU, memory, storage, egress, build minutes and spend. Select a project to see its units, volumes and buckets. |
Choose This month, Last month or Custom dates at the top. Days and months are in UTC. Export CSV downloads the breakdown by project and unit, with quantities and spend per resource.
What you pay for#
| Resource | Measured as | Priced per |
|---|---|---|
| CPU | The CPU each running unit reserves (its request), every hour it runs | vCPU-hour |
| Memory | The memory each running unit reserves | GiB-hour |
| Volumes | The size of each volume, from the moment it is created until it is deleted | GiB-month (730 hours) |
| Buckets | What is stored in each bucket, measured every 30 seconds | GiB-month |
| Egress | Bytes sent to the internet: responses the edge serves for your units, and connections your units open to public addresses | GiB |
| Builds | Minutes from a build's start to its end | build minute |
CPU and memory are billed on what a unit reserves, not on what it happens to use, because a reservation is what holds room for it on a node. The page shows what was used next to it, so you can see when a unit reserves far more than it needs and lower its request. Traffic between your own units, to private addresses and to the edge is not egress. A managed database is a unit, so it is billed for its CPU, memory and volume like any other.
Each hour is priced at the rates in effect when the hour started. If a price changes, the hours before the change keep their old price, so a past month never changes. The page notes any change that falls inside the period you are looking at.
Who sees amounts#
| Role | Sees |
|---|---|
| Owner and admin | Everything, including prices and dollar amounts. |
| API tokens | Everything (a token acts as an admin). |
| Member | Every quantity: CPU, memory, storage, egress and build minutes, by project and unit. Prices and amounts are hidden. |
How fresh it is#
Usage is gathered into hours every five minutes, so the current hour fills in as it goes. Nodes report CPU and memory every 15 seconds and keep their reports if they lose contact, so a node that was cut off catches up when it reconnects. Hours already counted are never counted twice.
Running your own Kuiper#
The platform admin sets prices on the admin dashboard's Pricing page and sees spend across every organisation on Billing. The operations guide in the Kuiper repository (docs/operations.md, under Prices and billing) lists the default prices and how they were chosen.