resemble-ai/chatterbox

Made with ♥️ by

Chatterbox is a family of state-of-the-art, open-source text-to-speech models by Resemble AI.

Chatterbox Multilingual V3 is the latest general-purpose multilingual TTS model in the Chatterbox family. It keeps the same 0.5B model size while improving speaker similarity, reducing hallucinations, and producing more natural, conversational speech across languages.

V3 is designed for broad language coverage like V2, but with stronger stability and more expressive generation. It is the recommended multilingual model for users who want one voice cloning model that works across many languages.

Alongside V3, we are releasing the Single Language Pack: dedicated finetunes for priority languages where tighter quality control, stronger language-specific behavior, and more specialized speech generation are valuable.

• Broad Multilingual Coverage: Designed as the main general-purpose multilingual Chatterbox model, supporting wide language coverage similar to V2.
• Single Language Pack: Dedicated single-language models provide stronger specialization and quality control where language- and regional-dialect-specific performance matters most.
• More Consistent Speaker Similarity: Improves voice identity and accent preservation across languages, making cross-language voice cloning more stable and reliable.
• Reduced Hallucination: V3 is optimized to reduce unwanted continuation, repetition, and off-prompt speech, especially in cases where earlier multilingual models were less stable.

For low-latency English voice agents, Chatterbox-Turbo is our most efficient model. Built on a streamlined 350M parameter architecture, Turbo delivers high-quality speech with less compute and VRAM than our previous models. We have also distilled the speech-token-to-mel decoder, previously a bottleneck, reducing generation from 10 steps to just one, while retaining high-fidelity audio output.

Paralinguistic tags are now native to the Turbo model, allowing you to use [cough], [laugh], [chuckle], and more to add distinct realism. While Turbo was built primarily for low-latency voice agents, it excels at narration and creative workflows.

For the most resource-constrained deployments, Chatterbox-Nano shares Turbo's architecture in an even smaller 110M parameter package. It targets on-device and CPU inference — running 3x faster than realtime on 8 CPU cores — while keeping the same single-step decoder and native paralinguistic tag support. Nano is the recommended model when memory and latency budgets are tightest.

If you like the model but need to scale or tune it for higher accuracy, check out our competitively priced TTS service (link). It delivers reliable performance with ultra-low latency of sub 200ms—ideal for production use in agents, applications, or interactive media.

Choose the right model for your application.

pip install chatterbox-tts

Alternatively, you can install from source:

# conda create -yn chatterbox python=3.11
# conda activate chatterbox

git clone https://github.com/resemble-ai/chatterbox.git
cd chatterbox
pip install -e .

We developed and tested Chatterbox on Python 3.11 on Debian 11 OS; the versions of the dependencies are pinned in pyproject.toml to ensure consistency. You can modify the code or dependencies in this installation mode.

import torchaudio as ta
import torch
from chatterbox.tts_turbo import ChatterboxTurboTTS

# Load the Turbo model
model = ChatterboxTurboTTS.from_pretrained(device="cuda")

# Generate with Paralinguistic Tags
text = "Hi there, Sarah here from MochaFone calling you back [chuckle], have you got one minute to chat about the billing issue?"

# Generate audio (requires a reference clip for voice cloning)
wav = model.generate(text, audio_prompt_path="your_10s_ref_clip.wav")

ta.save("test-turbo.wav", wav, model.sr)

Nano shares Turbo's architecture and is loaded through the same ChatterboxTurboTTS class by passing nano=True:

import torchaudio as ta
import torch
from chatterbox.tts_turbo import ChatterboxTurboTTS

# Load the Nano model (also runs on CPU: device="cpu")
model = ChatterboxTurboTTS.from_pretrained(device="cuda", nano=True)

# Generate with Paralinguistic Tags
text = "Hi there, Sarah here from MochaFone calling you back [chuckle], have you got one minute to chat about the billing issue?"

# Generate audio (requires a reference clip for voice cloning)
wav = model.generate(text, audio_prompt_path="your_10s_ref_clip.wav")

ta.save("test-nano.wav", wav, model.sr)

import torchaudio as ta
from chatterbox.tts import ChatterboxTTS
from chatterbox.mtl_tts import ChatterboxMultilingualTTS

device = "cuda"  # or "cpu" / "mps"

# English example
model = ChatterboxTTS.from_pretrained(device=device)

text = "Ezreal and Jinx teamed up with Ahri, Yasuo, and Teemo to take down the enemy's Nexus in an epic late-game pentakill."
wav = model.generate(text)
ta.save("test-english.wav", wav, model.sr)

# Multilingual V3 examples
multilingual_model = ChatterboxMultilingualTTS.from_pretrained(device=device, t3_model="v3")
# To use the legacy V2 multilingual checkpoint, omit t3_model or pass t3_model="v2".

french_text = "Bonjour, comment ça va? Ceci est le modèle de synthèse vocale multilingue Chatterbox, il prend en charge 23 langues."
wav_french = multilingual_model.generate(french_text, language_id="fr")
ta.save("test-french.wav", wav_french, multilingual_model.sr)

chinese_text = "你好，今天天气真不错，希望你有一个愉快的周末。"
wav_chinese = multilingual_model.generate(chinese_text, language_id="zh")
ta.save("test-chinese.wav", wav_chinese, multilingual_model.sr)

# If you want to synthesize with a different voice, specify the audio prompt
AUDIO_PROMPT_PATH = "YOUR_FILE.wav"
wav = model.generate(text, audio_prompt_path=AUDIO_PROMPT_PATH)
ta.save("test-2.wav", wav, model.sr)

See example_tts.py, example_tts_turbo.py, example_tts_nano.py, and example_vc.py for more examples.

The general-purpose Chatterbox Multilingual model supports the following languages:

Arabic (ar) • Danish (da) • German (de) • Greek (el) • English (en) • Spanish (es) • Finnish (fi) • French (fr) • Hebrew (he) • Hindi (hi) • Italian (it) • Japanese (ja) • Korean (ko) • Malay (ms) • Dutch (nl) • Norwegian (no) • Polish (pl) • Portuguese (pt) • Russian (ru) • Swedish (sv) • Swahili (sw) • Turkish (tr) • Chinese (zh)

The Single Language Pack provides dedicated finetunes for priority languages and regional variants. Use these when you want stronger language-specific behavior, tighter quality control, or dialect-aware generation beyond the general multilingual model.

• General Use (TTS and Voice Agents):

  • Ensure that the reference clip matches the specified language tag. Otherwise, language transfer outputs may inherit the accent of the reference clip’s language. To mitigate this, set cfg_weight to 0.
  • The default settings (exaggeration=0.5, cfg_weight=0.5) work well for most prompts across all languages.
  • If the reference speaker has a fast speaking style, lowering cfg_weight to around 0.3 can improve pacing.

• Expressive or Dramatic Speech:

  • Try lower cfg_weight values (e.g. ~0.3) and increase exaggeration to around 0.7 or higher.
  • Higher exaggeration tends to speed up speech; reducing cfg_weight helps compensate with slower, more deliberate pacing.

Every audio file generated by Chatterbox includes Resemble AI's Perth (Perceptual Threshold) Watermarker - imperceptible neural watermarks that survive MP3 compression, audio editing, and common manipulations while maintaining nearly 100% detection accuracy.

You can look for the watermark using the following script.

import perth
import librosa

AUDIO_PATH = "YOUR_FILE.wav"

# Load the watermarked audio
watermarked_audio, sr = librosa.load(AUDIO_PATH, sr=None)

# Initialize watermarker (same as used for embedding)
watermarker = perth.PerthImplicitWatermarker()

# Extract watermark
watermark = watermarker.get_watermark(watermarked_audio, sample_rate=sr)
print(f"Extracted watermark: {watermark}")
# Output: 0.0 (no watermark) or 1.0 (watermarked)

👋 Join us on Discord and let's build something awesome together!

Chatterbox Turbo was evaluated using Podonos, a platform for reproducible subjective speech evaluation.

We compared Chatterbox Turbo to competitive TTS systems using Podonos' standardized evaluation suite, focusing on overall preference, naturalness, and expressiveness.

Evaluation reports:

• Chatterbox Turbo vs ElevenLabs Turbo v2.5
• Chatterbox Turbo vs Cartesia Sonic 3
• Chatterbox Turbo vs VibeVoice 7B

These evaluations were conducted under identical conditions and are publicly accessible via Podonos.

• Podonos — for supporting reproducible subjective speech evaluation
• Cosyvoice
• Real-Time-Voice-Cloning
• HiFT-GAN
• Llama 3
• S3Tokenizer

If you find this model useful, please consider citing.

@misc{chatterboxtts2025,
  author       = {{Resemble AI}},
  title        = {{Chatterbox-TTS}},
  year         = {2025},
  howpublished = {\url{https://github.com/resemble-ai/chatterbox}},
  note         = {GitHub repository}
}

Don't use this model to do bad things. Prompts are sourced from freely available data on the internet.