Fish Audio releases Drama 3 emotive TTS
Drama 3 is an expressive text-to-speech model in 83 languages that you direct with plain-language cues. It is available in the Fish Audio app and API, and as open weights under a community license.
What it does
Drama 3 is Fish Audio's most emotive text-to-speech model. You direct it in square brackets the way you'd direct an actor, and can change direction partway through a line; an <emphasis>…</emphasis> tag directs only the words inside it.
It speaks 83 languages and is built for audiobooks, audio drama, game and animation dialogue, and short video.
Results and limits
Drama 3 is #1 on Artificial Analysis' Pronunciation Robustness benchmark, reading 93.2% of hard words, numbers and codes correctly, against 88.1% for S2.1 Pro. Other claims of "more feeling" come from Fish Audio's own demos.
Worth knowing
It's available on fish.audio, in the iOS and Android apps, and in the API (model header drama-3). Until November 12 the API is $10 per 1M characters (80% off the $50 list price). The weights are on Hugging Face under the Fish Audio Community License: free commercial use under US$500,000 annual revenue, with a separate license needed to offer it as a service.
Try it yourself
[barely holding it together, forcing a smile] [Line one]. [shaky breath] [Line two]. [whispering] [Final line].