Fish Audio: Fish Audio S2.1-Pro

Name: Fish Audio S2.1-Pro
Author: Fish Audio

fish/s2-1-pro

active tts 1 provider

Playground

Fish Audio's recommended production text-to-speech model (S2.1-Pro) with improved quality, latency and throughput over S2-Pro, plus a free variant (s2.1-pro-free) under fair-use limits.

AuthorFish Audio

Languages1 languages

VoicesLoading...

Response formatsMP3 · WAV · PCM · Opus

Variant models2 IDs

Statusactive

Quick start

Fish Audio S2.1-Pro

Create an API key

Create API key

Make a request

Pick an SDK, request mode, and language. The model and provider route are already filled in.

ElevenLabs SDK streaming TTS

Routes the ElevenLabs SDK through allmodels and streams the audio response.

import { ElevenLabsClient } from "elevenlabs";

const client = new ElevenLabsClient({
  apiKey: process.env.ALLMODELS_API_KEY,
  baseUrl: "https://api.allmodels.io/el"
});

const stream = await client.textToSpeech.stream("voice_id", {
  text: "I'm sorry, Dave. I'm afraid I can't do that.",
  modelId: "fish/s2-1-pro",
  outputFormat: "mp3_44100_128"
}, {
  queryParams: {
    provider: {
      "only": [
        "fish"
      ]
    }
  }
});

const reader = stream.getReader();
for (;;) {
  const { done, value } = await reader.read();
  if (done) break;
  // Write value to your response, file, or audio playback pipeline.
}

Pass provider parameters

Pass provider preferences through the compatible SDK request options when you want to pin a route to a provider.

{
  queryParams: {
    provider: {
      only: ["fish"]
    }
  }
}

Variant Models

These are native model IDs accepted by the author/provider for the same canonical model family.

Fish Audio S2.1-Pro Default

Recommended production Fish Audio TTS model; improved quality, latency, and throughput over S2-Pro.

s2.1-pro

Fish Audio S2.1-Pro Free Variant

Same model as S2.1-Pro billed at $0 for testing/dev/small business under fair-use limits.

s2.1-pro-free

Providers

Different providers can host the same model. Choose one provider when you need a fixed backend, or let routing select among them.

ProviderSlugPer 1k charactersLatencyiModesUptime

Fish Audio 19 params

fish

From $0.015

—

Synchronous · Streaming

Fish Audio

19 documented provider parameters for this route.

Pricing

ModeLanguagePer 1k characters

Text To Speech (Paid) Any $0.015 ($15.00 usd Per 1m Utf8 Bytes)

Free Tier Any $0 ($0.00 usd Per 1m Utf8 Bytes)

Parameters

model header · string · required

Specify which TTS model to use. We recommend `s2-pro`.

default "s2-pro"2 enum values

Values

s1s2-pro

text body · string · required

Text to convert to speech.

chunk_length body · integer · optional

Text segment size for processing.

default 300min 100max 300

condition_on_previous_chunks body · boolean · optional

Use previous audio as context for voice consistency.

default true

early_stop_threshold body · number · optional

Early stopping threshold for batch processing.

default 1min 0max 1

format body · string · optional

Output audio format.

default "mp3"4 enum values

Values

wavpcmmp3opus

latency body · string · optional

Latency-quality trade-off. normal: best quality, balanced: reduced latency, low: lowest latency.

default "normal"3 enum values

Values

lownormalbalanced

max_new_tokens body · integer · optional

Maximum audio tokens to generate per text chunk.

default 1024

min_chunk_length body · integer · optional

Minimum characters before splitting into a new chunk.

default 50min 0max 100

mp3_bitrate body · integer · optional

MP3 bitrate in kbps. Only applies when format is mp3.

default 1283 enum values

Values

64128192

normalize body · boolean · optional

Normalizes text for English and Chinese, improving stability for numbers.

default true

opus_bitrate body · integer · optional

Opus bitrate in bps. -1000 for automatic. Only applies when format is opus.

default -10005 enum values

Values

-100024000320004800064000

prosody body · object | null · optional

Speed and volume adjustments for the output.

default null

reference_id body · string | array | null · optional

Voice model ID(s) from Fish Audio library or your custom models. For single-speaker synthesis, provide a string. For multi-speaker synthesis (dialogue), provide an array of model IDs. **Multi-speaker is only available with the S2-Pro model.** When using multiple speakers, use speaker tags in your text like `<|speaker:0|>` and `<|speaker:1|>` to indicate speaker changes. Example: `<|speaker:0|>Hello!<|speaker:1|>Hi there!<|speaker:0|>How are you?` with `reference_id: ["speaker-a-id", "speaker-b-id"]`.

default null

references body · array | array | null · optional

Inline voice references for zero-shot cloning. Requires MessagePack (not JSON). For single speaker, provide an array of ReferenceAudio objects. For multiple speakers, provide an array of arrays where each inner array contains references for one speaker. **Multi-speaker is only available with the S2-Pro model.** The speaker index corresponds to the index in reference_id array. Example for multi-speaker: [[{audio, text}], [{audio, text}, {audio, text}]] for 2 speakers where speaker 1 has 2 reference samples.

repetition_penalty body · number · optional

Penalty for repeating audio patterns. Values above 1.0 reduce repetition.

default 1.2

sample_rate body · integer | null · optional

Audio sample rate in Hz. When null, uses the format's default (44100 Hz for most formats, 48000 Hz for opus).

default null

temperature body · number · optional

Controls expressiveness. Higher is more varied, lower is more consistent.

default 0.7min 0max 1

top_p body · number · optional

Controls diversity via nucleus sampling.

default 0.7min 0max 1

Supported languages

1 languages

English en

Voices

Features

Text to speech

Generates spoken audio from text input.

text_to_speech

Streaming

Supports streaming audio output (HTTP + WebSocket) with optional timestamps.

streaming

Synchronous

Synthesizes speech from fully-submitted text in a single synchronous request.

synchronous

Multilingual

Supports 83 languages.

multilingual

Voice cloning

Supports instant (reference audio) and persistent (reference_id) voice cloning.

voice_cloning

Emotion / expression control

Natural-language and bracket expression control; multi-speaker.

emotion_control

Code

ElevenLabs SDK streaming TTS

Routes the ElevenLabs SDK through allmodels and streams the audio response.

import { ElevenLabsClient } from "elevenlabs";

const client = new ElevenLabsClient({
  apiKey: process.env.ALLMODELS_API_KEY,
  baseUrl: "https://api.allmodels.io/el"
});

const stream = await client.textToSpeech.stream("voice_id", {
  text: "I'm sorry, Dave. I'm afraid I can't do that.",
  modelId: "fish/s2-1-pro",
  outputFormat: "mp3_44100_128"
}, {
  queryParams: {
    provider: {
      "only": [
        "fish"
      ]
    }
  }
});

const reader = stream.getReader();
for (;;) {
  const { done, value } = await reader.read();
  if (done) break;
  // Write value to your response, file, or audio playback pipeline.
}