Fish Speech is a multilingual text-to-speech and voice cloning model that supports over 80 languages and multi-speaker generation. It accepts text and short reference audio to generate speech audio without requiring phoneme preprocessing.
Project overview
Fish Speech is a multilingual text-to-speech and voice cloning model that supports over 80 languages and multi-speaker generation. It accepts text and short reference audio to generate speech audio without requiring phoneme preprocessing.
Project type
Model Runtime · Audio & Speech
Use cases
Audio & Speech
Deployment
Refer to project documentation
License
License pending
Best for
Developers, creators, and AI engineers requiring a multilingual TTS system that supports inline emotional tagging and multi-speaker generation.
Key capabilities
Generates natural and emotionally rich speech in over 80 languages without requiring phonemes or language-specific preprocessing.
Clones voices using short reference samples, typically 10 to 30 seconds, capturing timbre, speaking style, and emotional tendencies without additional fine-tuning.
Embeds emotional instructions at any position in the text using a simple tag syntax, such as [whisper], [excited], or [angry].
Processes reference audio containing multiple speakers and controls performance via speaker ID tokens, allowing a single generation to include multiple speakers.
Leverages previous information to improve the expressiveness of subsequent generated content, increasing the naturalness of dialogue.
Limitations and risks
The project is released under the FISH AUDIO RESEARCH LICENSE, which may restrict commercial usage.
The developers do not hold responsibility for illegal usage of the codebase and advise users to refer to local laws regarding the Digital Millennium Copyright Act (DMCA).
Getting started
Setup is rated as medium difficulty, requiring a Python environment, downloading large model weights, and configuring the inference server or WebUI. Users should follow the official installation instructions and run command line or WebUI inference.
Evidence and sources
GitHub project description: SOTA Open Source TTS
README: State-of-the-art multilingual text-to-speech (TTS) system, redefining the boundaries of voice generation.
README: On Seed-TTS Eval, S2 achieves the lowest WER among all evaluated models including closed-source systems
README: S2 Pro supports over 80 languages without requiring phonemes or language-specific preprocessing
README: Fish Audio S2 supports accurate voice cloning using short reference samples (typically 10-30 seconds). The model captures timbre, speaking style, and emotional tendencies, produci…