Usage
A project’s AI units, requests, and tokens over time, by model and by caller.
Read usage
/v1/projects/{project}/usageNeeds View usage and limits.
| Query | Takes |
|---|---|
range | 24h (the default), by hour; or 7d or 30d, by day |
{
"range": "7d",
"step": "day",
"current": {
"since": "2026-09-23T00:00:00Z",
"until": "2026-09-30T00:00:00Z",
"totals": {
"units": 4200,
"requests": 70,
"inputTokens": 91000,
"outputTokens": 30000
},
"points": [
{
"start": "2026-09-23T00:00:00Z",
"units": 600,
"requests": 10,
"inputTokens": 13000,
"outputTokens": 4300
}
]
},
"previous": {
"since": "2026-09-16T00:00:00Z",
"until": "2026-09-23T00:00:00Z",
"totals": {
"units": 3100,
"requests": 52,
"inputTokens": 70000,
"outputTokens": 21000
},
"points": []
},
"models": [{ "name": "openai/gpt-oss-120b", "units": 4200, "requests": 70 }],
"callers": [
{
"name": "apiKey:01m3c081d2wj4g63qcr53pa1yh",
"units": 4200,
"requests": 70
}
],
"tasks": [{ "name": "support-reply", "units": 900, "requests": 30 }]
}stepstringhourorday: how long each point is. Hours and days are in UTC.currentobjectThe time asked for, up to the end of this hour or today.
pointshas one entry for each step, in order, with zeros where nothing ran.previousobjectThe same length of time just before, to compare with.
current.totals.unitsintegerThe AI units calls used.
current.totals.requestsintegerCalls that used AI units.
current.totals.inputTokensintegerTokens sent to models.
current.totals.outputTokensintegerTokens models wrote.
modelsarrayThe current time’s units and requests by model, most first, up to 20.
callersarrayThe same by who called:
apiKey:and the key’s ID,web:and the web key’s ID for runs from web pages, oruser:and the person’s ID for calls from the console. Up to 20.tasksarrayThe same by task, for runs of tasks only. Up to 20.
A call counts in the hour it started.
Last updated