Skip to content
SPECIFICTION
Get listed

Chatterbox TTS

github.com

Free plan Open source WindowsmacOSLinuxAPISelf-hosted
Visit github.com
Spec score 9.1 #1 of 25 in Voice Cloning Software

What it is

Chatterbox TTS is a family of open-source text-to-speech models from Resemble AI. Its listed choices include Turbo, Nano, Multilingual V3, and dedicated single-language models. Turbo supports tags that can prompt reactions such as sighs, gasps, coughs, and laughter; Nano is intended for on-device or CPU use when memory and latency budgets are tight. The README lists 23 languages for Multilingual. The project supports voice cloning, audio input, batch processing, voice isolation, and WAV export. Generated audio carries Resemble AI’s PerTh watermark. The repository is MIT licensed, and its listed plan costs nothing for research and development. Resemble says it has stopped offering voice generation as a commercial product, while its models remain open and free for research and development. The README provides pip and source installation instructions and describes Python 3.11 on Debian 11, with CUDA, CPU, and MPS device options. Listed platforms are Linux, macOS, and self-hosted.

Spec sheet

What it does
Chatterbox is Resemble AI’s open-source family of text-to-speech and voice-cloning models.Read 21 Oct 2026 · raw.githubusercontent.com
Product
Chatterbox is a family of open-source text-to-speech models by Resemble AI.Read 8 Oct 2026 · github.com
Intended users
The README describes Turbo for low-latency voice agents and Nano for on-device or CPU inference when memory and latency budgets are tight.Read 8 Oct 2026 · github.com
Platforms
linux,macos,self_hostedRead 20 Sep 2026 · github.com
Free plan
YesRead 20 Sep 2026 · github.com
Support
The README invites users to join the project’s Discord community.Read 8 Oct 2026 · github.com
License
The code and model listing identify the project as MIT licensed, permitting modification and redistribution subject to the license notice.Read 21 Oct 2026 · github.com
Founded
2019Read 28 Sep 2026 · github.com
Commercial use
YesRead 20 Sep 2026 · github.com
Company founding
Resemble says Zohaib Ahmed and Saqib Muhammad started the company in 2019.Read 21 Oct 2026 · resemble.ai
Company headquarters
LinkedIn lists Resemble AI headquarters in Santa Clara, California, with 11–50 employees.Read 21 Oct 2026 · linkedin.com
Company ownership and funding
Crunchbase describes Resemble AI as a private, for-profit company with five funding rounds and founders Zohaib Ahmed and Saqib Muhammad, while funding amounts are obscured.Read 21 Oct 2026 · crunchbase.com
Controls
The Python interface exposes generation controls including exaggeration, CFG weight, temperature and random seed.Read 21 Oct 2026 · raw.githubusercontent.com
Current availability
Resemble says it has stopped offering voice generation as a commercial product and that its voice models remain open and free for research and development.Read 8 Oct 2026 · resemble.ai
Documentation depth
The repository includes model selection tables, Python examples, language identifiers, watermark extraction code, tuning guidance and evaluation links.Read 21 Oct 2026 · raw.githubusercontent.com
Export formats
WAVRead 20 Sep 2026 · github.com
Expressive speech
Turbo supports paralinguistic tags such as [sigh], [gasp], [cough], and [laugh] to prompt vocal reactions.Read 8 Oct 2026 · resemble.ai
Hallucination reduction
The V3 description says it reduces unwanted continuation, repetition and off-prompt speech compared with earlier multilingual versions.Read 21 Oct 2026 · raw.githubusercontent.com
Installation
The project can be installed with pip as chatterbox-tts or from a cloned repository in editable mode.Read 21 Oct 2026 · ondalabs.app
Integration
NVIDIA provides a Chatterbox TTS plugin for ACE for Games with CUDA, Vulkan and Windows D3D12 backends.Read 21 Oct 2026 · docs.nvidia.com
Languages
The README lists 23 supported languages for Chatterbox Multilingual.Read 8 Oct 2026 · github.com
Latest model
The repository identifies Chatterbox Multilingual V3 as its newest general-purpose multilingual model.Read 21 Oct 2026 · raw.githubusercontent.com
Limit: commercial availability
Resemble’s current product notice says voice generation is no longer offered as a commercial product, although the models remain open and free for research and development.Read 21 Oct 2026 · resemble.ai
Limit: comparative performance evidence
An independent review reports that Resemble’s blind test favored Chatterbox, but an Artificial Analysis snapshot ranked Chatterbox below ElevenLabs’ leading model.Read 21 Oct 2026 · ondalabs.app
Limit: service reliability and billing
Trustpilot complaints repeatedly mention failed or cut-off audio, slow or absent support, unexpected charges, credit problems and service outages; these reviews concern Resemble’s former hosted service rather than the self-hosted model.Read 21 Oct 2026 · trustpilot.com
Limit: watermark removability
An independent review notes that MIT licensing allows modified software, so users can technically remove the watermark by changing the code.Read 21 Oct 2026 · ondalabs.app
Maker legal name
Resemble AI, Inc. is the company name shown in the maker’s privacy-policy contact information.Read 21 Oct 2026 · resemble.ai
Memory requirement
NVIDIA documents approximately 2.3 GB VRAM for Turbo-only inference and 3.5 GB for multilingual TTS-only inference in its plugin.Read 21 Oct 2026 · docs.nvidia.com
Model choices
The project lists Chatterbox Turbo, Nano, Multilingual V3, and dedicated single-language models.Read 8 Oct 2026 · github.com
Multilingual support
The multilingual model lists 23 languages, including Arabic, Chinese, English, French, German, Hindi, Japanese, Korean, Spanish and Turkish.Read 21 Oct 2026 · raw.githubusercontent.com
Nano model
Chatterbox-Nano is a 110-million-parameter model intended for CPU and on-device inference, with the repository claiming three-times-realtime speed on eight CPU cores.Read 21 Oct 2026 · raw.githubusercontent.com
Paralinguistic tags
Turbo supports tags such as [laugh], [cough] and [chuckle] inside text prompts.Read 21 Oct 2026 · raw.githubusercontent.com
Praise: voice quality and usefulness
A minority of Trustpilot reviewers praised the voices as useful for marketing and said customer support resolved an accidental subscription issue.Read 21 Oct 2026 · trustpilot.com
Public API
Replicate exposes an official Resemble Chatterbox model through an API with prompt, reference-audio, temperature, CFG and exaggeration inputs.Read 21 Oct 2026 · replicate.com
Rating on Gartner Peer Insights
No rating; the Resemble AI Platform page says there are no reviews yet.Read 21 Oct 2026 · gartner.com
Rating on Trustpilot
1.8 out of 5 from 27 reviews.Read 21 Oct 2026 · trustpilot.com
Replicate price
Replicate lists the hosted Chatterbox model at $0.025 per 1,000 input characters.Read 21 Oct 2026 · replicate.com
Runtime
The README says Chatterbox was developed and tested with Python 3.11 on Debian 11 and shows device options including CUDA, CPU, and MPS.Read 8 Oct 2026 · github.com
Security and compliance
Resemble AI’s Trust Center states that it is ISO/IEC 27001:2022 certified and has completed a SOC 2 Type II examination.Read 8 Oct 2026 · trust.resemble.ai
Self-hosting
Chatterbox is run locally and does not require an account when installed from the open-source project, according to an independent technical guide.Read 21 Oct 2026 · ondalabs.app
Single-language models
The project publishes dedicated 500-million-parameter language packs for Chinese, several Spanish and Portuguese variants, and Hindi.Read 21 Oct 2026 · raw.githubusercontent.com
Supported devices
The README examples support CUDA, CPU and Apple MPS device selection.Read 21 Oct 2026 · raw.githubusercontent.com
System requirement
NVIDIA’s Chatterbox plugin documentation specifies Windows 10 or 11 64-bit, 16 GB RAM and an NVIDIA GPU with compute capability 7.0 or newer for that integration.Read 21 Oct 2026 · docs.nvidia.com
Turbo model
Chatterbox-Turbo is a 350-million-parameter English model designed for lower compute and latency.Read 21 Oct 2026 · raw.githubusercontent.com
Voice cloning
Models can generate speech conditioned on a short reference audio clip.Read 21 Oct 2026 · raw.githubusercontent.com
Voice similarity
V3 is described as improving speaker identity and accent preservation across languages.Read 21 Oct 2026 · raw.githubusercontent.com
Watermarking
Every generated audio file receives Resemble’s PerTh imperceptible neural watermark, which the project says survives MP3 compression and common editing.Read 21 Oct 2026 · raw.githubusercontent.com

Plans and prices

Chatterbox open-source models Free Free to use under the MIT license Open-source models for research and development Read 8 Oct 2026 · github.com
Voice AI Research (Chatterbox) Free Open models · free to build on · no commercial hosted voice service Read 21 Oct 2026 · docs.resemble.ai

Every Chatterbox TTS plan and price →

Strong points and limits

Strong points

  • MIT-licensed models are listed at free.
  • Turbo supports prompted vocal reactions.
  • Multilingual lists 23 supported languages.
  • WAV export and voice cloning are supported.

Limits

  • Generated audio includes a PerTh watermark.
  • Resemble no longer offers voice generation commercially.
  • Listed platforms are Linux, macOS, and self-hosted.

Who it suits

It suits developers building voice agents or exploring text-to-speech and voice cloning with open-source models. Nano is aimed at on-device or CPU inference under tight memory and latency budgets.

Verdict

Chatterbox offers several open-source model options, including multilingual and low-latency uses. Buyers should account for the audio watermark and the research-and-development availability described by Resemble.

Where Specifiction places it

Alternatives to Chatterbox TTSAll alternatives →