@rohanpaul_ai: Really cool, open weight, TTS model, just dropped. Fixing one bad word in a TTS clip usually means regenerating the who…

X AI KOLs Timeline Models
text-to-speech open-weight voice-synthesis emotional-tts fish-audio

Summary

Fish Audio 发布了开源权重的 TTS 模型 Drama 3，号称全球最具情感表现力和可控性的语音合成模型，支持在句子中对单个单词进行风格和语气微调。

Really cool, open weight, TTS model, just dropped. Fixing one bad word in a TTS clip usually means regenerating the whole line, and the take you liked goes with it. @FishAudio just released Drama 3, it's open-weight, and the most emotive and controllable TTS model in the world > lets users restyle single words mid-sentence > Directions go inside square brackets in plain language, such as “barely holding it together,” and they can be written in English or the user’s own language. > paired tags such as Whisper, Slower or Forceful wrap only the selected words, so a line can turn from steady to hushed halfway through.
Original Article
View Cached Full Text

Cached at: 10/14/26, 01:59 PM

Really cool, open weight, TTS model, just dropped.

Fixing one bad word in a TTS clip usually means regenerating the whole line, and the take you liked goes with it.

@FishAudio just released Drama 3, it’s open-weight, and the most emotive and controllable TTS model in the world

lets users restyle single words mid-sentence

Directions go inside square brackets in plain language, such as “barely holding it together,” and they can be written in English or the user’s own language.

paired tags such as Whisper, Slower or Forceful wrap only the selected words, so a line can turn from steady to hushed halfway through.

Fish Audio (@FishAudio): $52M seed round. 15M+ users. But this is what we’re most proud of yet.

Introducing Drama 3: the most emotive and controllable TTS model in the world

And yes, it’s open-weight.

You can now control how even a single word should sound, change its tone, and shift your voice

Similar Articles

Fish Audio Drama 3

Product Hunt

Fish Audio 推出其最新 AI 语音模型 Drama 3，支持通过自然语言控制情感、节奏、耳语、笑声和停顿，可生成多角色场景、有声书与游戏对话，覆盖 83 种语言，并以社区许可证开放权重供自托管。

fishaudio/drama-3

Hugging Face Models Trending

Fish Audio 发布新一代文本转语音模型 Drama 3，采用 35B 参数 MoE 架构（每 token 激活 3B），支持最多 16 个说话人的多角色长文本生成、语音克隆与自然语言控制，并可在 SGLang-Omni 上流式部署。