Fish Audio Drama 3
Product Hunt Justin Jincaid ● Covered by 2 sources
Fish Audio launched Drama 3, an AI voice model you can direct with plain language. It can shift emotion, pacing, whispers, laughter, and pauses in one sentence.
Based on reporting by Product Hunt, Justin Jincaid — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Fish Audio has launched Drama 3, its latest AI voice model, and the pitch is clearer than most text-to-speech hype: this is for directing performances, not just generating voice clips.
The model lets users steer emotion, pacing, whispers, laughter, and pauses with natural language, even inside a single sentence. That makes it a fit for multi-character scenes, audiobooks, game dialogue, and narration, where flat delivery usually gives the game away.
Fish Audio says Drama 3 supports 83 languages and is available on the web, mobile, and API. It also comes with open weights for self-hosting under Fish Audio's community license, so teams can run it themselves instead of treating it like a black box.
This is Fish Audio's fifth launch, and the company is leaning hard into expressive voice work. The broader pitch is simple: more control, more nuance, less robotic delivery. In voice AI, that still matters more than another generic talking machine.
My take — AI-written commentary, not fact-checked reporting
This is the right direction for voice AI: less “press generate” and more actual direction. Closed demos are fine for applause, but open weights and self-hosting are where serious creators and serious paranoia both end up. The awkward truth is that the best synthetic voices will keep getting more human, while the product copy keeps pretending that’s the cute part.
Read more about this at: Product Hunt