GitHub radar

YuE2 music generation now in ComfyUI

Comfy-Org packaged the m-a-p YuE2-3B music generation model as a native ComfyUI node — describe a mood and genre, get a track, right inside your existing workflow.

01Comfy-Org/YuE2 15379k downloads/mo3B paramstext-to-audio

Comfy-Org packaged the YuE2-3B music generation model from m-a-p into ComfyUI format, making it a native workflow node. YuE2 generates full tracks — melody, rhythm, instrumentation — from text descriptions of genre and mood. The package has two parts: the SheetSage2 audio encoder and the generation model itself. Installation follows standard ComfyUI practice: one file in the checkpoints folder, one in audio_encoders. The model runs under a cc-by-nc-4.0 license, so commercial use is off the table. Community demos show it working for ambient, lo-fi, and pop styles. Quality matches a compact 3B model — fine for background tracks and content work, not professional music production.

Why a vibe-coder should care

If you already generate images or video in ComfyUI, you can now add music generation to the same workspace without switching to a separate tool. Useful for content creators who need background tracks for reels or short-form video. It runs locally — no Suno account, no monthly subscription, no API keys.

How to install

Copy this and send it to your agent — Claude Code, Codex, any of them:

Set up the YuE2 music generation model for ComfyUI from Comfy-Org: https://huggingface.co/Comfy-Org/YuE2 — download both files, place them in the right folders, and show me how to open an example workflow and generate my first track.

A regular laptop with 16 GB of RAM — the bf16 model takes about 6 GB.

Open on Hugging Face