Fish Audio API (S2.1 Pro)
Multilingual (83 languages per third parties) cloning-first TTS with a MessagePack WebSocket for streamed text. Billed per UTF-8 byte, with a free s2.1-pro-free model under fair use.
Overview
Best for: Multilingual cloning and character voices at a low flat price.
At a glance
$15 per 1M UTF-8 bytes, so non-Latin scripts cost about 3x per character. Latency (~90 ms) and 83 languages come from third-party coverage. Free model s2.1-pro-free has no SLA, is time-limited and may retain data. Open weights (S2 Pro) are non-commercial. Concurrency 5 until $100 paid.
Text (MessagePack frames on WebSocket)
mp3 (default, 64/128/192 kbps), wav, pcm, opus; 44.1 kHz for most formats, 48 kHz for opus
83 per third-party coverage of S2.1 Pro (not on the docs pages checked)
Very large community voice library; instant cloning via references
Vendor claim (third-party reported): ~90 ms to first audio for S2.1 Pro; Vapi measured 141 ms median including network.
Not specified
Not verified.
Features
- input streaming (MessagePack WebSocket)
- latency modes low/balanced/normal
- voice cloning from reference audio
- multi-speaker dialogue
- prosody speed/volume
Pricing
| What | Price | Unit |
|---|---|---|
| s2.1-pro / s2-pro / s1 | $15.00 | per 1M UTF-8 bytes |
| s2.1-pro-free | $0.00 | per 1M UTF-8 bytes |
| voice-design-1 | $0.01 | per successful request |
900 chars/min. English ASCII = 1 byte/char ($0.0135/min); CJK/Arabic/Hindi are ~3 bytes/char (~$0.04/min). Free model excluded.
Free tier: s2.1-pro-free model at $0 under fair use (time-limited per third-party sources)
Source: docs.fish.audio
Setup
- Create an API key and prepay credit at fish.audio.
- Choose a voice reference_id (or upload references).
- Open wss://api.fish.audio/v1/tts/live with Authorization: Bearer and a model header.
- Send start, text..., flush, stop as MessagePack; collect 'audio' events until 'finish'.
- Or use the Python SDK's stream_websocket with a text generator.
Endpoint
wss://api.fish.audio/v1/tts/liveAuthentication
Authorization: Bearer <FISH_API_KEY>; header model: s2.1-pro
Quick start python
# Fish Audio Python SDK (import name 'fishaudio'; check PyPI for the package name)
from fishaudio import FishAudio
client = FishAudio() # reads FISH_API_KEY
def llm_tokens():
for token in ["The ", "first ", "move ", "sets ", "everything ", "in ", "motion."]:
yield token
with open("out.mp3", "wb") as f:
for chunk in client.tts.stream_websocket(llm_tokens(), reference_id="YOUR_VOICE_ID"):
f.write(chunk) # play or forward as it arrivesWritten from the current docs. Check the vendor's SDK version before you ship.
Warnings
Billed per UTF-8 byte, not character
Non-Latin scripts cost ~3x per character (Chinese, Japanese, Korean, Arabic, Hindi are 3-4 bytes each). Compare vendors on bytes for non-English workloads.
Free model is temporary and may train on your data
s2.1-pro-free has no SLA, is reported to run only through 2026-11-30, and Fish's blog says requests may be retained for model improvement. Do not send sensitive text.
s1 retires 2026-12-31
After that date s1 requests are served by s2.1-pro; voices may sound different.
Low default concurrency
Only 5 concurrent requests until you have paid $100; 429s carry no Retry-After header.
Cloning community voices
The public voice library contains user-uploaded clones; confirm you have rights to any voice you use commercially.
Open weights are non-commercial
The S2 Pro weights on Hugging Face use the Fish Audio Research License, not a commercial licence; self-hosting for a product needs a separate agreement.
Plus 12 warnings that apply to all text-to-speech, streaming APIs. See category warnings.
Limits
- Concurrent requests: Starter (<$100 paid) 5, Elevated ($100+) 15, High Volume ($1,000+) 50
- 429 without Retry-After; use exponential backoff
- chunk_length 100-300
Models and products
| Name | Status |
|---|---|
| s2.1-pro | GA (default if model header omitted) |
| s2.1-pro-free | Free, fair use, no SLA |
| s2-pro | GA |
| s1 | Deprecated, retiring 2026-12-31 |
Docs and sources
Docs
Sources used
- docs.fish.audio/developer-guide/models-pricing/pricing-and-rate-limits.md
- docs.fish.audio/api-reference/endpoint/websocket/tts-live.md
- fish.audio/ko/blog/s2-1-pro-free-api/
- humannessindex.vapi.ai/models/fish-s2-1-pro
- huggingface.co/fishaudio/s2-pro
Language count (83) and ~90 ms latency come from third-party coverage; end date of the free model; Python package name.