Audio length
Longer audio uses more credits. Test with short text to keep it cheap.
One credit pool covers AI voice generation (text-to-speech and voice cloning). Cost is shown before each run, and failed jobs are never charged.
What's included
What's included
What's included
One-time top-ups — buy more credits whenever you run low.
What's included
What's included
What's included
Powered by
Use a single balance across the supported voice and audio models, and check the credit cost before each request.
audio models
Model credit guide
You're only charged when a generation succeeds. The exact estimate in the generator depends on the model and the length of the audio you generate.
| Type | Model | Credit cost |
|---|---|---|
| Voice | Seed Audio 1.0 | Text-to-speech (TTS). Credit cost scales with the length of the generated audio (input text), with a minimum floor, and is always shown before you generate. |
| Voice cloning | Seed Audio 1.0 · Clone | Generated from reference audio (up to 3 clips, 30s each). Cost is length-based, same as standard text-to-speech. |
Credits scale with the length and settings of the audio you generate.
Longer audio uses more credits. Test with short text to keep it cheap.
Cloning from reference audio may use slightly more than plain TTS.
Sample rate and output format can affect usage.
The estimated cost is shown before each generation.
Common questions about credits and voice generation cost.
New accounts get free credits to try voice generation. After that, top up via subscription or credit packs.
By the length and settings of the audio you generate. The cost is shown before each run.
No. Failed jobs are not charged and reserved credits are refunded.
Please check the terms for commercial use.
Enter text and check the cost before you generate.