Remy Reference/models/Chatterbox Multilingual

Chatterbox Multilingual

Chatterbox Multilingual, a text-to-speech model from DeepInfra.

speechModelOverride: { model: "chatterbox-multilingual" }

RResemble AI·served viaDeepInfra·Released 2025

Pricing

List price: $1 / 1M characters
Per 1M characters: $1.00
Per 1M characters: $0

Specifications

Context window: 2,000 tokens
Released: 2025
Model string: ResembleAI/chatterbox-multilingual
Capabilities: Open sourceLow costMultilingual

Parameters

Model options, passed in speechModelOverride.config.

Parameter	Type	Default	Range / options
`language_id` The language the text is in.	Select	en	ardadeelenes+17
`exaggeration` How expressive the delivery is. Higher is more dramatic; pair it with a lower guidance for a slower, more deliberate read.	Number	0.5	0–1, step 0.05
`cfg` How closely the speech follows the text and the voice.	Number	0.5	0–1, step 0.05
`temperature` How much the delivery varies from one reading to the next.	Number	0.8	0.05–2, step 0.05
`seed`	Seed	—	—
`response_format`	Select	mp3	mp3wavflacopus

Use this model

await mindstudio.textToSpeech({
  text: "...",
  speechModelOverride: {
    model: "chatterbox-multilingual",
    config: {
      language_id: "en",
      exaggeration: 0.5,
    },
  },
});