DiffSinger

Design by Marble

Logo made by Marbles

What is DiffSinger?

DiffSinger is an AI-based diffusion singing voice synthesizer that is open source, originally developed by MoonInTheRiver, and is superceeded by the OpenVPI team. OpenUtau is currently used as a frontend for DiffSinger, until OpenVPI finishes their work on the GUI.

Tutorial

coming soon...

Vocal downloads:

Japanese

No7 (SEVEN)

Namine Ritsu

Deshimaru Gakuto

PJS

Otozora Rinly

Printto Magicbeat Japanese

Kurobousoku

Hikuine Kiiro

Chinese

ZhiBin

灰雨千

English

Peiton DS

Tiger

POTG

Russian

Keiro
(JP also supported)

Monone
(EN/JP also supported)

Ashera Lyre
(EN/JP also supported)


Llane Crow

Brazilian Portugese

Umidaji Rainbow-α

Polish

PIX AI v2.0
(JP also supported)

Thai

Printto Magicbeat Thai