API Reference
Log In
API Reference

Get cost per customer over time

Returns the cost in USD of the AI calls made for each customer, per time bucket, over the date range. The bucket size grows with the length of the range. The default is the sum; use metricType for the average, maximum and so on. Use customers to keep only some customers and costSource to choose which cost to count. For one figure per group over the whole period, a far smaller answer over a long date range, use get_customer_cost_aggregated, which does not take costSource, customers.

Recent Requests
Log in to see full request history
TimeStatusUser Agent
Retrieving recent requests…
LoadingLoading…
Query Params
date-time

Start of the date range, ISO 8601 in UTC, for example 2026-01-01T00:00:00Z. Defaults to 24 hours before endDate.

date-time

End of the date range, ISO 8601 in UTC. Defaults to now.

string
enum

How the values in each group are combined: sum, avg, max, min, count, median, or the percentile p50, p90, p95 or p99. Defaults to sum.

integer | null
≥ 0
Defaults to 0

Zero-based page of _embedded.items to return. Defaults to 0.

integer
≤ 100
Defaults to 20

Number of _embedded.items per page, from 1 to 100. Defaults to 20.

customers
array of strings

Keep only these customers, by name. Repeat the parameter to pass several values.

customers
costSource
array of strings

Count only cost from these sources: revenium_metered (usage metered by the Revenium SDK) or provider_billing (cost reported by your AI provider's billing). Omit it to count both. Repeat the parameter to pass several values.

costSource
Allowed:
string

Hashed id of the team to answer for. Optional. With a user token it may be any team the user belongs to; without it the user's default team answers. An API key always answers for its own team, and naming any other team is refused with 403.

Headers
string
enum
Defaults to application/json

Generated from available response content types

Allowed:
Responses

Language
Credentials
Bearer
API Key
LoadingLoading…
Response
Click Try It! to start a request and see the response here! Or choose an example:
application/json
application/problem+json