Simple pricing, billed per second
Pay only for the audio you transcribe or produce and the minutes your voice agents are connected. No subscription, no minimum per request, no charge for failed requests.
| Service | Mode | Per hour | Per minute |
|---|---|---|---|
| Speech-to-TextRealtime (streaming) | Realtime (streaming) | CHF 2.20 | CHF 0.037 |
| Speech-to-TextBatch (files and jobs) | Batch (files and jobs) | CHF 1.35 | CHF 0.023 |
| Text-to-SpeechRealtime (streaming) | Realtime (streaming) | CHF 6.85 | CHF 0.114 |
| Text-to-SpeechBatch (requests and jobs) | Batch (requests and jobs) | CHF 4.60 | CHF 0.077 |
| Voice agents (speech-to-speech)Realtime (streaming) | Realtime (streaming) | CHF 19.80 | CHF 0.33 |
Prices in CHF, excluding VAT. Metered per second: speech-to-text and text-to-speech by the audio you send or we produce, voice agents by connected session time, from the start of the session until it closes.
60 free minutes
Every new account starts with free credit that covers an hour of realtime text-to-speech, to spend on any service at its own rate: more minutes of speech-to-text, fewer of voice agents. No expiry date, no credit card required.
Get 60 free minutesFree sandbox keys
Sandbox keys return realistic test audio and transcripts with the exact formats and headers of the real service. They are never billed, which makes them ideal for automated tests.
Read the docs →Enterprise
Large volumes, dedicated capacity, contractual availability and latency, a data processing agreement and invoicing. We will put together an offer for your use case.
Talk to usEstimate your monthly cost
Prices in CHF, excluding VAT. Metered per second: speech-to-text and text-to-speech by the audio you send or we produce, voice agents by connected session time, from the start of the session until it closes.
How billing works
Prepaid credit
You top up credit, online with TWINT or card, or by invoice for companies. Every service draws on the same balance at its own rate. Credit does not expire.
Per second, no minimum
A two-second prompt costs two seconds. There is no rounding up to a minute and no charge for requests that fail.
Always visible
The console and the usage endpoint show what you have used, down to the second, before any invoice is issued.
Alerts before you run out
Everyone on your account is notified at 80 % and 95 % of the credit used. You can add alerts of your own.
Frequently asked questions
What counts as realtime and what as batch?
Realtime means the streaming WebSocket endpoints, where audio or text flows while it is being produced, and voice agents, which are always realtime. Batch means whole files and whole texts: single requests and jobs.
How are voice agents billed?
Per second of connected session time, from the moment the session is ready until it closes, at CHF 0.33 per minute (CHF 19.80 per hour) excluding VAT. It draws on the same prepaid credit as the other services. Sandbox sessions are never billed.
Do the free minutes or the credit expire?
No. Neither the 60 free minutes nor purchased credit have an expiry date. Both stay available for as long as your account exists.
Are failed requests billed?
No. Only audio that was actually transcribed or produced is metered. Sandbox traffic is never billed.
Is there a volume discount?
Yes, for committed volumes. Tell us about your use case and expected hours per month.
Which currency and VAT apply?
Prices are in Swiss francs and exclude VAT. Swiss VAT is added for customers in Switzerland and Liechtenstein.
Try it on your own audio
Get 60 free minutes of speech-to-text and text-to-speech. No credit card required.