Fish Audio

Fish Audio is a voice AI platform built for realistic text-to-speech, fast voice cloning, and accurate transcription. Creators, developers, and teams use it to turn text into natural sounding speech, clone a voice from a short recording, and convert speech into readable text.

The platform is powered by its own S2 and S2.1 Pro models, trained on millions of hours of audio across dozens of languages. This lets it produce speech with natural pacing, tone, and emotion, including inline tags for effects like whispering, laughing, or sighing.

Beyond text-to-speech, Fish Audio includes a voice changer, sound effect generator, audio separation tool, audio translation, and a Story Studio for combining audio into longer projects. A library of more than two million community voices is also searchable and ready to use.

Fish Audio is available as a web app for everyday use, an API for developers building apps and voice agents, and as open source models that can be self-hosted for research or production.

• Realistic text-to-speech with emotion tags

• Fast voice cloning from short samples

• 2 million plus searchable voice library

• Speech-to-text with speaker detection

• Voice changer for recorded audio

• Sound effect generation

• Audio separation into stems

• Audio translation and dubbing support

• Story Studio for longer audio projects

• Pay-as-you-go developer API