Video

Fish Audio

AI text-to-speech and voice cloning platform with emotion tags, a 2M+ voice library, speech-to-text, and a voice API

text to speechvoice cloningemotion controlaudiobooksvoice API

4.5 / 5Open Site

Fish Audio is a voice AI platform for text-to-speech, voice cloning, speech-to-text, and avatar lip-sync. You write or paste text, pick from a library of 2,000,000+ voices, and add inline emotion and special tags such as angry, whispering, laughing, sighing, or pause to direct the delivery. Voice cloning works from a short sample (the company says about 15 seconds), the platform supports 30+ languages with any voice, and Story Studio helps produce audiobooks with chapter-level control and audio that meets ACX and Audible specs. Other tools include a voice changer, audio translation, sound effects, audio separation, an end-to-end voice agent solution, an MCP connection, and an API with low-latency streaming for building products. An avatar lip-sync feature pairs generated images and videos with Fish Audio voices. Plans are credit-based, with a Free tier of 8,000 credits a month for personal, non-commercial use, and paid plans that allow commercial use of voices you own. The company, Hanabi AI Inc., reports 8M+ users and publishes its Fish Audio models and open-source Fish Speech research.

Best for

Creators, game and animation teams, audiobook producers, and developers who want expressive, emotion-controlled voices and voice cloning with a free tier and an API

Pricing

Free tier: 8,000 credits a month (up to 7 minutes, 500 characters per generation, personal use); Plus $15/mo (listed), first year 50% off at $66 billed annually, 250,000 credits and up to 200 minutes; Pro $100/mo, first year 50% off at $450 billed annually, 2,000,000 credits and 3 team seats; Max $999/mo, first year 50% off at $4,494 billed annually, 25,000,000 credits and 10 team seats; Enterprise custom; roughly 600 to 625 credits per minute