Simple, honest pricing
Every tier includes a monthly allowance of generated audio. Upgrade any time.
Free
₹0
Try it out, watermarked, real-time voices.
Premium
—/mo
Priority queue on dedicated GPUs, plus 250 minutes of credits.
Premium Plus
—/mo
Everything in Premium, our best models, and 800 minutes of credits.
Compare what each plan includes
Minutes are only part of the difference. Compare voice capacity, queue access and model quality before you choose.
| Feature | Free | Starter | Premium | Premium Plus |
|---|---|---|---|---|
| Monthly generated audio | ~180 seconds / month | ~60 minutes / month | ~250 minutes / month | ~800 minutes / month |
| Reusable cloned voices | 1 voices | 5 voices | 20 voices | 50 voices |
| Generation queue | Standard | Standard | Priority | Priority |
| Model access | Real-time | Creator quality | Creator quality | Flagship models |
| Maximum single generation | Up to 3 min | Up to 3 min | Up to 10 min | Up to 20 min |
| Saved history | Latest 2 | Latest 10 | Latest 10 | Latest 10 |
| Commercial usage | Not included | Included | Included | Included |
| AI provenance watermark | Included | Included | Included | Included |
Allowances are approximate because actual usage depends on the generated audio length and model. Translating a script for multi-language generation also draws on your minutes: about 1 second per language for every 1,000 characters. Current checkout pricing always comes from the published catalogue.
What uses your minutes
Your plan's minutes are spent only when we make new audio. The Generate page shows an estimate before you start.
- Generating speech
- Uses the length of the audio you get. Indian languages and some voice models use minutes faster than others.
- Generating in several languages
- Each language is its own recording, so three languages use about three times the minutes of one.
- Translating a script
- About 1 second per language for every 1,000 characters. Paid plans only.
Included, without using minutes: cloning and saving voices, writing and editing scripts, and replaying or downloading audio you have already made.
If a generation fails, it uses nothing.