Usage analytics
Analytics charts how a project uses the API: how many requests, what they cost, how many tokens, and which models and keys they came from.
What you can see
- Requests, Spend, Saved and Tokens totals for the selected range. Saved is the difference between list price and what you paid, from any discounts on your organization.
- A chart of Requests, Spend or Tokens over time, stacked by model or by API key, with a table of the values.
The six largest models or keys get their own series; the rest are grouped as "Other". Requests from the playground appear as "Playground", and keys that have since been deleted keep their history.
In a Token Plan workspace, Spend is shown as Credit consumed.
Filters
| Filter | Options |
|---|---|
| Project | Any project in the organization. |
| Date range | From the last 5 minutes to the last 365 days. |
| Breakdown | By model or by API key. |
The filters are kept in the URL, so you can bookmark a view or send it to a teammate.
New requests show up within about a minute.
The dashboard
The dashboard shows the headline numbers for the whole organization — total requests, spend and savings — plus the selected project's last seven days: requests per day, tokens, spend, error rate and top models.
Exact numbers
Analytics is for trends. For the cost and token counts of a single request, use the request logs; billing always follows the per-request figures.