Pricing and metering
How model choice, audio duration, processing stages, and volume shape what an Ultrafield request costs, and how usage is metered per second.
Ultrafield pricing is designed around the processing a request performs rather than a fixed bundle of unrelated features.
What shapes usage
- Speech model and batch or streaming path
- Processed audio duration
- Diarization or speaker-identification stages
- Formatting, entity, sentiment, and other understanding processors
- Hosted language-model tokens
- Upcoming third-party provider usage through the model gateway
Published per-minute rates for speech-to-text, diarization, and speaker identification are on the pricing page. Rates for the remaining processors are being finalized and are supplied with API access.
Volume planning
For sustained workloads, estimate monthly audio minutes, peak concurrency, typical file duration, model mix, and downstream processor mix. Ultrafield can use that profile to plan capacity and a more predictable commercial structure.
See the pricing page or contact us with a representative workload.