ToRouter DocsUsage details & CSV export
Filter request logs by date and model, break down tokens, and export everything to CSV.
/usage is the detailed log view. Use it for reconciliation, cost attribution, and exporting to your billing system.
Filters
- Date range — start / end date pickers; the chart and table both respect this
- Model — narrow to a single model id
- Key — when you have multiple API keys, scope to one
The top stat cards summarise the filtered window: total cost, total requests, input/output token counts.
Per-model breakdown
The model table shows, per model in the date range:
| Column | What |
|---|---|
| Model | Model id you called |
| Requests | Successful request count |
| Input tokens | Prompt tokens billed |
| Output tokens | Completion tokens billed |
| Cache creation | Anthropic prompt-cache write tokens (when applicable) |
| Cache read | Anthropic prompt-cache read tokens (cheaper) |
| Cost | USD spend |
Export to CSV
Click Export CSV in the top-right of the details table. ToRouter paginates the export in batches of 1000 rows server-side and assembles a single file named usage-<start>_<end>.csv. The file includes per-request rows, not just the model summary.
Exports can be large. Narrow the date range or model filter before exporting if you only need a slice — full-month dumps on busy accounts can run to tens of thousands of rows.
Request log entries
Each row in the log shows: timestamp, model, total tokens (with input/output/cache breakdown on hover), cost, and key name. Cache creation tokens split further into 5-minute and 1-hour TTL when the upstream returns that detail.
See also
If you suspect a shared API key is being used by someone else, sign in to the console: narrow this table with the key filter, review per-key stats on the API keys page, and cross-check model mix on the dashboard.