TTS Dashboard watcher, jobs, outputs, services

Experiment workspace

Voice Lab

Describe a narrator, generate a reference performance, then keep or discard the result.

Qwen3-TTS-12Hz-1.7B-VoiceDesign

01 · Describe

Design the voice

NATURAL LANGUAGE
VoiceDesign uses the description as its speaker instruction.

02 · Listen

Audio preview

completed
Test bd5323ff2026-09-04T06:57:21.221029+00:00

Replay the preview as you refine the description or reference text.

Describe→Generate→Listen→Keep or discard