Analyze Cost and Usage

The report needs traffic through at least one LLM provider. With no provider configured, the page shows a No providers configured notice with an Add an LLM provider button in place of the report. With a provider but no traffic in the window, the figures read zero.

Open Cost and usage in the sidebar. The page has two tabs:

The report header names its scope, All gateway traffic, and the selected range, next to the date-range picker and the Export menu. The footer repeats the range and notes that the figures are in USD, aggregated by day (or by hour for a window of 24 hours or less) in UTC, and cover model costs only.

The date-range picker offers last 7 days (the default), last 14 days, last 30 days, last 90 days, month to date, quarter to date, year to date, and a custom range. Every figure on the tab, including each comparison with the previous period, follows this one window.

When you open the report without a range in the URL, it starts on the default. If that range holds no traffic, the report checks the wider rolling presets and switches in one step to the smallest one that does, up to last 90 days. The picker always names the range actually in use. Choosing a range yourself, including a custom one, stops the switch for the rest of the visit, and it doesn’t add a browser history entry. The Activity tab widens its own window too, one preset at a time.

A custom range sets an exact start and end time, to the minute and interpreted in UTC, rather than a relative period. It writes customStart and customEnd ISO-8601 timestamps to the page URL, so the view is shareable. Copy the URL after picking a custom range, and any teammate who opens it lands on the same window.

A link from another page, such as the gateway traffic card on the home page, can open the report on a one-day window that the picker itself doesn’t offer. The report then uses hourly buckets, and the spend chart is titled Hourly spend instead of Daily spend. A custom range of 24 hours or less also uses hourly buckets.

Four figures summarize all gateway traffic in the window:

The previous period is the window of the same length that ends where the selected window starts. For a 7-day range, that is the 7 days before it. For a one-day range, it is the 24 hours before. The same comparison draws the dashed line in the spend chart and fills the Change column of the breakdown table. When the previous period had no spend, Total spend shows no percentage change.

The report has two spend charts:

To fill the resource chart, pick a resource with the Choose a resource picker on the Resource spend chart, or click Compare trend on a row of the breakdown table. The picker searches the agents, models, providers, and users with spend in the window. Clear selection empties the chart again. While a resource is charted, the panel header shows its spend and its share of gateway spend.

The two charts share a hover cursor. Point at a bucket in either chart and both highlight the same bucket. The spend chart’s tooltip shows this period’s value, the previous period’s value and date, and the change between them. The resource chart’s tooltip shows the resource’s value for that bucket.

The page URL records the charted resource, so a compared view is shareable. Copy the URL, and a teammate lands on the same chart. Tag values are the exception. You can chart one from its row’s details in the breakdown table, but you can’t pick it here, and it isn’t kept in the URL.

The Spend breakdown table lists every agent, model, provider, or user with spend in the window, sorted by spend. Switch the grouping with the table’s Agents, Models, Providers, and Users tabs. When cost-allocation tags are available, a Tags tab appears as well. Pick a tag key next to it to group by that key’s values. See Track Spend by Tag.

Grouping applies to this table only. The headline figures and the spend chart always cover all gateway traffic, and the resource chart keeps whatever you charted.

Each row shows the resource’s spend, its change against the previous period, its request count, and its cost per thousand requests. Click a column header to sort by it, and use the search box to narrow the rows by name. Some rows stand for spend that has no resource of that kind. Direct usage collects calls that didn’t run through an agent, Unattributed collects calls with no user recorded, and Untagged collects calls without a value for the chosen tag key. A user row that belongs to an agent’s service account says so under the name, and a provider that has since been deleted keeps its row, marked Deleted provider, so history stays complete. When the window holds more resources than the table loaded, the table says so and suggests a narrower range.

Click a row to see its details:

The Export menu in the report header offers two CSV downloads:

Two settings in the same menu shape the full report. Under Full report time bucket, choose Hourly, Daily (the default), Monthly, or Total (no time column). Under Full report rows, choose any combination of Provider, Model, User, and Agent to decide which columns the rows are grouped by. Clearing all four gives one row per time bucket.

Use the breakdown export to share the view you have on screen, and the full report when you need the whole dataset for a spreadsheet or a chargeback pipeline.

The Activity tab answers a different question from the report: instead of what something cost, it shows what’s running right now. It ranks the agents with traffic over a window of the last hour (the default), the last 24 hours, or the last 7 days, and shows each agent’s change against the previous window. Select an agent to see its recent requests and open a conversation.