resemble-ai/chatterbox-multilingual

OfficialView on Replicate →

Generate expressive, natural speech in 23 languages. Features instant voice cloning from short audio, emotion control, and seamless cross-language voice transfer.

Capabilities

Seed

Cost

Community model (estimated from hardware time)

Input Parameters

Name	Type	Description	Default	Constraints
`text`*	string	Text to synthesize into speech (maximum 300 characters)	`—`	—
`cfg_weight`	number	CFG/Pace weight controlling generation guidance (0.2-1.0). Use 0.5 for balanced results, 0 for language transfer	`0.5`	min: 0.2, max: 1
`exaggeration`	number	Controls speech expressiveness (0.25-2.0, neutral=0.5, extreme values may be unstable)	`0.5`	min: 0.25, max: 2
`language`	string	Language for synthesis	`"en"`	ardadeelenesfifrhehiitjakomsnlnoplptrusvswtrzh
`reference_audio`	string(uri)	Reference audio file for voice cloning (optional). If not provided, uses default voice for the selected language.	`—`	—
`seed`	integer	Random seed for reproducible results (0 for random generation)	`0`	—
`temperature`	number	Controls randomness in generation (0.05-5.0, higher=more varied)	`0.8`	min: 0.05, max: 5

textrequiredstring

Text to synthesize into speech (maximum 300 characters)

cfg_weightnumber

CFG/Pace weight controlling generation guidance (0.2-1.0). Use 0.5 for balanced results, 0 for language transfer

Default: 0.5min: 0.2, max: 1

exaggerationnumber

Controls speech expressiveness (0.25-2.0, neutral=0.5, extreme values may be unstable)

Default: 0.5min: 0.25, max: 2

languagestring

Language for synthesis

Default: "en"

ardadeelenesfifrhehiitjakomsnlnoplptrusvswtrzh

reference_audiostring

Reference audio file for voice cloning (optional). If not provided, uses default voice for the selected language.

seedinteger

Random seed for reproducible results (0 for random generation)

Default: 0

temperaturenumber

Controls randomness in generation (0.05-5.0, higher=more varied)

Default: 0.8min: 0.05, max: 5

Version: 9cfba4c265e6Updated: 7/25/202624.1K runs