Fish Audio

About

A voice AI platform for text-to-speech, speech-to-text, and voice cloning, built on its S2.1 Pro model. Scripts can be marked up with inline tags such as [whispering], [excited], or [laughing] so the delivery changes mid-sentence, which is what makes it usable for longer audiobook and character work rather than single lines. It clones a voice from roughly 15 seconds of audio, supports 30+ languages, and exposes the same models through APIs and SDKs for voice agents and chatbots.