Fish Audio S2.1 Pro
Fish Audio's recommended production TTS model, covering 83 languages with natural-language delivery control using bracket syntax and multi-speaker dialogue.
- Weights
- Not released
- API
- Yes
- Commercial use
- Permitted subject to conditions
Pricing is a flat $15 per 1M UTF-8 bytes, and a free variant of the same model is offered without latency or DPA guarantees. Automatic language detection is built in.
Voice cloning is referenced in the API but consent and commercial-use terms for clones were not stated on the docs pages fetched.
- Requires consent from the person depicted. Voice cloning is referenced in the API but consent and commercial-use terms for clones were not stated on the docs pages fetched
Specification
- What it takes in
- text, audio
- What it makes
- audio
- Controls
- style-prompting, emotion-control, multi-speaker-dialogue, voice-cloning, streaming-low-latency
- Weights
- Not released
- API
- Yes
- Pricing
- $15.00 per 1M UTF-8 bytes for s2.1-pro, s2-pro and s1 alike; s2.1-pro-free is $0.00. Docs note 1M UTF-8 bytes is roughly 180,000 English words or about 12 hours of speech. ASR (transcribe-1) $0.36 per audio hour; Voice Design $0.01 per successful request. Concurrency tiers by cumulative prepay: 5 requests under $100, 15 at $100+, 50 at $1,000+.
- Released or updated
- not established
- Languages
- 83, not individually listed by the vendor
Strengths
- 83 languages - the broadest verified count in this dataset apart from ElevenLabs v3
- Flat $15 per 1M UTF-8 bytes, roughly 12 hours of speech - by far the cheapest hosted expressive TTS here
- A genuinely free production-grade tier (s2.1-pro-free)
- Multi-speaker dialogue and bracket-syntax delivery control
- Predecessor s2-pro documents 100ms time-to-first-audio
Weaknesses
- No latency figure published for s2.1-pro itself (only for s2-pro)
- Sample rate and max text length not documented
- Free tier carries no TTFA or DPA guarantee
- Consent/licensing terms for cloned and community voices are unclear
- Low concurrency (5 requests) until you prepay $100+
Best for
- multilingual VO on a budget
- expressive dialogue scenes
- high-volume localisation drafts
Where the sources disagree
Recorded rather than resolved. Picking the tidier number would hide the fact that the vendor's own pages do not agree.
- Language counts differ by model in the vendor docs: s2.1-pro 83, s2-pro 80+, s1 13. A third-party review claimed '70+ languages' for S1, which contradicts the vendor's own 13.
- A third-party source dated S1's release to 29 July 2026, which conflicts with vendor docs describing S1 as the previous generation behind S2 and S2.1; the date is not recorded.
What to understand before using it
Explainers naming the mechanisms this model's own documentation mentions.
See it from the vendor
Shotcraft hosts no sample output. Judging a model means watching its own showreel with your own eyes, and a clip re-hosted here would be a clip we chose.
Nearby in the directory
Azure AI Speech text-to-speech (Neural, Neural HD, MAI voices)
Microsoft's enterprise speech synthesis service, with 200+ prebuilt neural voices across 80+ locales plus tiers for HD voices, multilingual voices, mu
Voice and speech
Sonic 3.5
Cartesia's current TTS model, documented at sub-90ms first-byte latency across 42 languages. It is tuned for conversational delivery and for reading a
Voice and speech
XTTS-v2
An open-weights multilingual TTS and voice cloning model covering 17 languages at 24kHz, still widely used for local voice cloning. Coqui Inc. shut do
Voice and speech