SGLang says Drama 3 from Fish Audio is supported on day 0 in SGLang-Omni, serving multi-speaker TTS, voice cloning, and streaming audio through /v1/audio/speech. It reports a median time to first audio of 89ms on a single H200 for a single stream. The image shows the command sgl-omni serve with --model-path fishaudio/drama-3 on port 8000.
They say seeing is believing. Wait till you hear what Drama 3 has to say. Day-0 support for @FishAudio Drama 3 is live in SGLang-Omni!
· Serve multi-speaker TTS, voice cloning, and streaming audio through /v1/audio/speech
· 89ms median time to first audio on a single H200 (single stream)
Full cookbook 👇: https://sgl-project.github.io/sglang-omni/cookbook/fishaudio_drama3.html
$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 mid-sentence. Uncle Roger was impressed too. Try it here: https://s.fish.audio/rqh0l8 Comment “DRAMA” within the next 48 hours for 1 month free. If you're an enterprise your first month is on us: https://s.fish.audio/qz9r0t
Source: SGLang · x.comPublished · added here
