音声
Chatterbox
Chatterbox — Resemble AI's open-source TTS family: expressive text-to-speech and voice cloning, multilingual V3, low-latency Turbo and tiny Nano models, MIT
★ 27.1k スターMIT
ガイドを見る →Fish Speechは、fishaudioが公開しているオープンソースの音声プロジェクトです（GitHub: fishaudio/fish-speech）。GitHubスターは33,289、フォークは2,891で、Custom (check repo)ライセンスの下で公開されています。GitHubでの最終更新は2026年10月14日です。
Fish Speech — Fish Audio's expressive open text-to-speech and voice cloning: SOTA-quality TTS you can self-host, under the Fish Audio Community License (check terms)
Follow the official installation guide (linked from the README at speech.fish.audio/install): environment setup and weight downloads, or use the Docker deployment the repo ships.
Get the current Fish Speech weights from the project Hugging Face organisation (fishaudio) as the install guide directs.
Run inference through the WebUI for interactive cloning/TTS, the command line for batches, or the API server when another app should call it.
Natural, expressive synthesis with voice cloning from reference audio — the project headline capability.
Three ways in: a browser UI for people, a CLI for scripts, and an API server for products.
The repo ships Dockerfiles and compose files for a self-hosted deployment without hand-building the environment.
The S2 model line and Hugging Face releases are current — the repo was pushed in October 2026 and its licence was refreshed the same month.
The README routes all setup through the official docs: environment, weights, Docker.
https://speech.fish.audio/install/The repo ships compose files (compose.yml and variants) for the container deployment described in the docs.
git clone https://github.com/fishaudio/fish-speech.git
cd fish-speechCLI, API server and WebUI inference are documented step by step here.
https://speech.fish.audio/inference/Fish Audio (fishaudio on GitHub), the team behind the fish.audio platform; this repo is their open TTS system.
For research and non-commercial use, yes, free of charge under the Fish Audio Community License. Limited commercial use is allowed below a revenue threshold; larger commercial or hosted-service use needs a separate licence from Fish Audio — read the LICENSE file in the repo.
Only clone voices you have the right to use — your own or with clear permission. The project disclaims responsibility for illegal usage, and consent and local law are on the user.
Against Chatterbox (also new on LibreAI this week), Fish Speech is the heavier, hosted-platform-grade system with a custom community licence; Chatterbox is MIT-licensed and simpler to embed. Against XTTS-style research models, Fish Speech is the more productised, actively released option.
Follow the install guide at speech.fish.audio/install (or the Docker route), download the weights from Hugging Face, then generate from the WebUI.
Chatterbox — Resemble AI's open-source TTS family: expressive text-to-speech and voice cloning, multilingual V3, low-latency Turbo and tiny Nano models, MIT
Self-host Chatterbox TTS with web UI, API, and voice cloning.
Easy-to-use Speech Toolkit including Self-Supervised Learning model, SOTA/Streaming ASR with punctuation, Streaming TTS with text…
IndexTTS: industrial-level, controllable zero-shot text-to-speech with voice cloning
MOSS-TTS family: open text-to-speech models for voice cloning, dialogue, voice design, and real-time agents
Open-source AI video localization and dubbing for YouTube/Bilibili: speech recognition, subtitle translation, voice cloning, audio mixing and rendering. 开源 A…
LibreAIカタログに掲載された音声プロジェクトの中で最もスターの多いもの — Fish Speechが合わなかった場合に検討したい、現実的な候補リストです。
利用 AI 大模型和自动化工作流，根据主题或关键词一键生成高清短视频。Generate HD short videos from a topic or keyword with an automated AI workflow.
Robust Speech Recognition via Large-Scale Weak Supervision
1 min voice data can also be used to train a good TTS model! (few shot voice cloning)
Clone a voice from 5 seconds of audio and generate arbitrary speech in real time
Port of OpenAI's Whisper model in C/C++
カタログの中でFish Speechとスター数が近い音声プロジェクト — 前後の棚もそのまま見て回れます。
🚀Clone a voice in 5 seconds to generate arbitrary speech in real-time
Chatterbox — Resemble AI's open-source TTS family: expressive text-to-speech and voice cloning, multilingual V3, low-latency Turbo and tiny Nano models, MIT
Instant voice cloning from MIT and MyShell — clone a voice from a short reference clip
Faster Whisper transcription with CTranslate2
VoxCPM2: tokenizer-free TTS for multilingual speech, creative voice design and true-to-life cloning
IndexTTS: industrial-level, controllable zero-shot text-to-speech with voice cloning