Pricing & metering
Usage-based, metered per call — the same shape as a modern API platform, with the unit adapted to video. You only pay for what you process.
How metering works
Billed per minute of video processed, or per request for lookups. Rounded up to the nearest second. Processing APIs (Shield, Intelligence, Transcription, Tagging, Model ID) bill per minute of video; lookups (Fingerprint match, §2257) bill per request.
If you’ve billed against the Claude API, the model is familiar: a metered per-unit rate, usage tiers, and committed-use discounts — the unit here is a minute of video instead of a token.
What you’re billed on
Under the hood these are GPU inference jobs, so compute is the real cost — but you’re never billed in raw GPU-seconds. The price is a stable proxy: a minute of video. Each API carries a different per-minute rate because each uses a different amount of GPU (a transcription pass is light; a full intelligence pass is heavy), so compute cost is baked into the rate, not exposed as a meter. Your cost is predictable before you make the call.
Per-call rates
Illustrative list prices, pending finance sign-off. Test-mode (jw_test_) keys
are intended to be non-billable.
Video-length passes (Shield, Scene Intelligence, Transcription, Model ID) bill per minute. We’re reviewing whether discrete, request-shaped operations — Tagging and Fingerprint · index — should bill per request instead, matching §2257 lookup and Fingerprint · match. Rates here are not final.
Volume tiers
Rates step down automatically as monthly volume grows; the top band is a committed-use annual contract.
Included with Ultimate
Every Ultimate scene includes its Shield scan at no extra metered cost. Standalone PAYG is for teams that want a single API against content they host themselves — available on any plan, billed only by usage.
Rate limits ≠ quota
- Rate limit — requests per minute (concurrency). Hit it and you get a
429withRetry-After. See Errors & limits. - Quota / metering — your dollar usage. Prepay with credits + auto-reload, or settle monthly. Set a budget alert so spend never surprises you.
Track usage
Live usage, per-API breakdown, and budget alerts live in Settings → Developers. Every metered call also appears on your monthly invoice, grouped by API and site.