Fish Audio pricing
What Fish Audio charges per million characters, what that costs at real volumes, and when a cheaper or pay-once option makes sense.
Prices are list prices from the official pricing page of Fish Audio, read October 8, 2026. They change; confirm before you decide. Source: https://docs.fish.audio/developer-guide/models-pricing/pricing-and-rate-limits
What Fish Audio charges
| Item | Price | Unit | Notes |
|---|---|---|---|
| Text to speech, s2.1-pro | $15 | per 1M UTF-8 bytes | English is about one byte per character; other scripts use more bytes per character |
| Voice design | $0.01 | per request | |
| Speech to text | $0.36 | per hour | |
| Voice cloning | No separate price listed on the page we read | - | Reference audio can be sent inline or saved as a voice model |
Cost at real volumes
| Model | 1M characters | 10M characters | 100M characters |
|---|---|---|---|
| s2.1-pro text to speech | $15 | $150 | $1,500 |
| Speakvora catalog voices | $3 | $30 | $300 |
One million characters is very roughly 15 to 20 hours of speech. Cost only: these models sound different, and we do not claim ours match theirs. Listen to the free samples and decide.
What Fish Audio offers
- Text to speech with the s2 model family and multi-speaker dialogue
- Cloning from reference audio
- Speech to text
What we did not find in Fish Audio's official material
- Curated industry phrase packs
- Downloadable ready-generated industry audio
- A listed price for custom-voice generation separate from the base rate
How it compares with Speakvora
Fish Audio bills text to speech by UTF-8 bytes at $15 per million and focuses on expressive models and cloning. Speakvora's catalog voices are $3 per million characters, and its own-voice rate is $6; we make no claim about Fish Audio's voice quality.
Pay once instead
If most of your audio repeats (greetings, confirmations, prompts), a one-time bundle plays offline with no per-play fee. See pay-once text to speech and the speech API alternative page.
More on Fish Audio: · · cheapest text to speech API · .
Questions
How much does Fish Audio cost?
For text to speech, Fish Audio lists between $15 and $15 per million characters depending on the model, on its pricing page read October 8, 2026. Prices change, so confirm at the source: https://docs.fish.audio/developer-guide/models-pricing/pricing-and-rate-limits.
Is there a cheaper alternative to Fish Audio?
On list price per character, yes. Speakvora's catalog voices are $3 per million characters and your own voice is $6. That is a cost comparison only; the models, languages and voices differ, and we do not claim equal quality. Fish Audio bills text to speech by UTF-8 bytes at $15 per million and focuses on expressive models and cloning. Speakvora's catalog voices are $3 per million characters, and its own-voice rate is $6; we make no claim about Fish Audio's voic
What does Fish Audio do that Speakvora does not?
Fish Audio offers: Text to speech with the s2 model family and multi-speaker dialogue; Cloning from reference audio; Speech to text. Speakvora is smaller, with fewer models and languages.
Can I pay once instead of per character?
For lines that repeat, yes. Speakvora sells ready-made voice bundles you buy once (audio from $14, or $214 to $399 with a local agent) and play offline with no per-play fee. If you generate a line once and keep the file, per-character pricing is already very cheap, and we say so on the comparison pages.