A voice generation environment unifying Qwen3-TTS + Irodori-TTS

Kotonia Voice is a free voice synthesis workspace bundling Apache 2.0 / MIT OSS TTS models. Switch between pro character speakers, voice cloning from a reference, and text-described voice design from a single screen.

10 langMultilingual narration

3sClone reference audio

FreeLocal GPU free queue

High-quality TTS in 10 languages

Natural narration in 10 languages including JP, EN, ZH, and KO via Qwen3-TTS's 9 brand speakers.

Zero-shot voice cloning

Pass a 3-second reference and transcript to read any text in that voice.

JP voice design

Generate voices from captions like "low-voiced woman, holding back irritation" with Irodori-TTS.

Fully in-browser

Chunked streaming starts playback during generation; download the full WAV when done.

Short-form video narrationCharacter voice prototypesMultilingual content dubbingVoice sample collection / A-B comparison

Standard textQwen3-TTS CustomVoice. 9 brand speakers x 10 languages, with instruct-based tone control.

Speaker (9 voices)

Language

Instruct (optional, tone control)

Speed (0.7-1.3)1.00×

Output

After generating you can play or download here.

History (latest 20)

No generation history yet.