Every AI call on ToRun writes exactly one billing record at the moment the call executes. The price snapshot is frozen at that instant, so your ledger stays verifiable even after providers change their rates. Here is how to read it.

What each record shows

Open Billing from the sidebar and select any row. You will see:

• Model and provider — which model handled the call and which provider served it.
• Capability cost breakdown — costs are split by pricing unit (input tokens, output tokens, cached input tokens, images, audio seconds, and so on). If a single call used vision and text together, you see each component separately.
• Price snapshot — the per-unit rate ToRun applied at execution time, quoted in USD. This rate is immutable; it will not change retroactively.
• Currency snapshot — the canonical USD amount plus, if you pay in a local currency, the local amount, the FX rate, and the timestamp when that rate was captured.
• Fund source — whether the call drew from your free session budget, a subscription allowance, your wallet balance, or a BYOK-routed account.
• Source tag — what triggered the call (a chat message, a workflow step, an image generation, and so on).

How daily summaries roll up

Your Usage dashboard aggregates billing records into daily summaries. Each summary covers one calendar day in UTC and rolls up total spend, token volume, and call count. The summaries update in near-real time as calls land — there is no end-of-day batch delay.

Exporting for your own audit

1. Go to Billing > Export.
2. Choose a date range (up to 90 days per export).
3. Select CSV or JSON.
4. Download the file. Every row maps one-to-one to a billing record, so you can recompute any total offline using the frozen price snapshots in each row.

For current rate information, see the pricing page.