Download
Kaizen Speech Studio Kaizen Speech Studio Help All Help Contact
Reference

Cost management (Azure)

Speech Studio doesn't charge you — Azure does. Here's how to keep Azure spend predictable.

Who charges whom

  • StepForward Solutions / Kaizen Apps charges you for the Speech Studio license (one-time or yearly).
  • Microsoft Azure charges you for API usage (TTS minutes, transcription minutes, video dubbing, storage).

Free tier (F0)

Azure Speech Service's F0 pricing tier includes 500,000 characters of standard Neural TTS per month (≈9 hours) at zero cost, renewing every month. Speech Studio includes no speech minutes of its own — the Free download carries a $1 test credit. On Pro with your own key, Microsoft’s F0 quota is yours.

The free hours belong to the resource's tier, not to your Azure account. You only get them if the Speech resource was created on F0. A resource on S0 has no free allowance and bills every character. Check: Azure Portal → your Speech resource → Overview → Pricing tier. If it says S0, follow the fix in the setup guide.

Other F0 facts worth knowing:

  • One F0 Speech resource per Azure subscription.
  • Standard Neural voices only. Neural HD, MAI-Voice, OpenAI and custom voices are not part of the free allowance.
  • 20 requests per minute and up to 10 minutes of audio per request.
  • An S0 resource can't be downgraded to F0 — create a new resource on F0 instead.

The $200 sign-up credit is not the free tier

New Azure accounts get a one-time credit (currently $200) that expires after 30 days. On your invoice it appears as a Credits line that cancels out your Charges. It applies to any tier, so an S0 resource looks free for the first month and then starts billing the day the credit runs out. The F0 allowance is different: it never expires and resets every month.

Paid tier (S0) — pay-as-you-go

On an S0 resource Azure charges per unit from the first character. Approximate pricing (check Microsoft for current rates):

  • Text-to-Speech: ~$15 per million characters for standard Neural voices. Neural HD and MAI-Voice are billed at higher per-character rates — see Azure Speech pricing.
  • Speech-to-Text: ~$1 per hour
  • Video Translation: varies, check Video Translation pricing
  • Blob Storage: ~$0.018 per GB/month + transaction costs (typically cents per month)

How to monitor spend

  1. Azure Portal → your subscription → Cost Management + Billing.
  2. Set a monthly budget alert (e.g. $10).
  3. Check weekly for the first month to calibrate.

Tips to stay cheap

  • Use the $1 test credit in the Free download for experimentation.
  • Switch to your own key once workload justifies it.
  • For long-form audio, generate once and save. Don't re-run the same chapter unless the content changed.
  • HD and MAI-Voice voices are billed at a higher rate than standard Neural and aren't covered by F0. Use standard Neural voices when you want to stay at $0.
  • Video translation is the most expensive feature per minute. Test on short clips first.

Reading your Azure invoice

  • AI + Machine Learning is the category Speech usage falls under, together with any other Azure AI service you use.
  • Charges is what your usage cost at list price. Credits is sign-up credit applied against it. Total is what your card is charged.
  • An F0 resource never produces TTS charges within its quota. If Speech charges appear every month, the resource is on S0.

Speech Studio's cost estimates

The History panel shows an estimated cost per run based on Azure's published prices. Treat it as a guideline; actual Azure billing is authoritative.