Usage dashboard

The Usage page (admin only) is accessible from the sidebar. It shows aggregated statistics for all requests proxied through the LLM gateway.

LLM Usage dashboard showing summary cards, request and cost charts, and per-instance and per-provider breakdowns

It shows:

Summary cards — total requests, input tokens, cached input tokens, output tokens, and estimated cost for the selected period
Requests over time — area chart of daily request volume
Cost over time — area chart of daily spend in USD
Cost by instance — horizontal bar chart showing which instances are spending the most
Cost by provider — horizontal bar chart breaking down spend by provider
Tokens by model — stacked bar chart of the top 10 models by token volume

Use the date range pickers and dropdowns at the top of the page to filter by time period, instance, or provider.

The Reset button permanently deletes all usage log records. This action cannot be undone.

The cost is calculated based on our catalog of well-known providers