Skip to content
Hangry LabsUseful tools, easy to run
Overview Listen to examples Documentation GitHub ↗

YuE2 listening room

From words to music.

Original lyrics, a little musical direction, and a local GPU. Listen to the results, explore the recipes, and try a prompt on your own machine.

8 steps48 kHz stereoGenerated locally

About these recordingsDownload all prompts ↗

Hangry Labs YuE2

Loading recordings…

Choose a song to listen.

Make it your own

Start with a prompt. Find your sound.

  1. Download a song's Prompt JSON.
  2. In your local YuE2 studio, select Open and choose the file.
  3. Edit the lyrics or musical direction. Leave seeds blank for a fresh take, then generate.
Run the local studio ↗

These songs use original demo lyrics and text prompts, with no uploaded reference recordings. Genre and voice descriptions are model hints. Listening examples do not establish benchmark results.

The publisher lists English and Chinese. Japanese and Spanish requests here are experimental. Full-song pronunciation and lyric fidelity have not been reviewed by fluent listeners. Interface language never changes the original lyrics, direction or recordings.

Upstream model card · Language notes

Model terms apply separately from the application's MIT license. Detailed technical references currently remain in English.

Model sources and terms · Download provenance manifest · Generation guide

Hangry Labs Local tools. Your machine. Your work. View source