Fish Audio Raises $50M to Scale Open-Source Speech Models with Fine-Grained Controls

Fish Audio, a Palo Alto startup that lets developers steer AI voices with over 15,000 natural language commands, has raised $50 million in seed funding. The round was led by Coreline Ventures and Capital Today, with participation from a roster of investors including 359 Capital, Parable, and HF0. The company says it now serves 8 million users across its open-source and hosted model versions, and has reached $21 million in annual recurring revenue since launching last year.

The startup began as a side project from former NVIDIA researcher Shijia Liao, who trained a voice generation model on a single GPU and open-sourced it. That repository, Fish Speech, now has more than 31,000 GitHub stars and is used by indie developers, game designers, and creators. Fish Audio has since released five models—four for speech generation and one for speech-to-text—with three of the generation models fully open-source. Its latest, S2.1 Pro, is available only through a paid API. The company offers monthly plans for creators and teams, plus an enterprise tier used by companies like HeyGen, Sanas, and Plaud.

To build its voice library, Fish Audio lets users submit their own voices for training and compensates them if their voice is used. That approach drew controversy months ago when some creators said their voices appeared without consent. CEO Rissa Cao told TechCrunch the company now has an automated takedown process: creators can prove ownership with a short voice sample or contract, and the voice is removed in under three minutes. Still, the system relies on creators discovering unauthorized uploads. Coreline Ventures partner Oskue Honda stressed that trust is fragile: "Consent, transparency, and attribution must be built into the product." Looking ahead, Fish Audio plans to release an audio understanding model and a speech-to-speech model. With big competitors like ElevenLabs and Cartesia in the race, investor Rico Mallozzi of 359 Capital says fine-grained controls and cost-efficient training give Fish Audio an edge.

Source: TechCrunch