Get usage overview
Your organization’s usage over a time window, bucketed for charting.
One request serves a whole dashboard: totals for the window, plus the series
behind them at the requested granularity. Every bucket in the window is
returned, including the empty ones, oldest first — so a chart can be drawn
straight from buckets with no gaps and no sorting. The totals are the sum of the
same buckets, so the two always agree.
Covers the whole organization, including rooms a teammate ran.
The window may cover at most 90 days, counted as calendar days in the
requested timezone; a longer window returns 422. The response is not paginated —
request a shorter window or a coarser granularity for a smaller series.
Remaining credit is not part of this: a balance is a level rather than a total over
a window, so it lives on GET /v1/billing/credit.
Authorizations
Bearer authentication header of the form Bearer <token>, where <token> is your auth token.
Query Parameters
Start of the window, inclusive. Must carry a UTC offset, e.g. 2026-07-30T00:00:00Z. With ends_before it may cover at most 90 days.
End of the window, exclusive: an event at exactly this instant belongs to the next window, so consecutive windows tile without double-counting. Must carry a UTC offset, and must be after starts_after.
Bucket width, which also sets how many points come back — the series is dense, so a 90-day window is 13 weekly points or roughly 2,160 hourly ones.
hour, day, week IANA zone name the buckets are truncated in, so a day starts at local midnight and a week on the local Monday. A fixed offset such as +05:30 is not accepted, because an offset cannot express DST.
Response
Successful Response
Usage totals and the series behind them, for one time window.
The requested window, granularity and timezone are echoed back, so the response
describes exactly what it covers. buckets is oldest first and includes empty
buckets, so it can be charted without further processing.
How wide each bucket of a usage series is.
hour, day, week Totals for the whole window, summed from the buckets below them.
total_participants counts the people who connected — a participant who drops and
rejoins counts twice, and the translator serving each room is not counted, though
its minutes are included in total_minutes.