Song Description

Random example

Inspiration

Save to...

What Is YuE 2?

YuE 2 is the open-weights song model for lyrics-first creation. You bring the full lyric sheet and a style prompt — no AI ghostwriter — then tune prompt adherence on a 0–20 scale and optionally reference an uploaded track for a music-to-music pass. One MP3 song per run, with the full editing suite enabled.

  • Free to Start

    Sign up and get free credits to test every model — no subscription required.

  • Commercial Ready

    Every song you generate is licensed for commercial use at no extra cost.

What Is YuE 2?

What YuE Does Best

Literal lyrics control and tunable style fidelity.

Your words, kept whole

YuE never rewrites your input: the lyrics you paste are the lyrics it sings, line by line. The most literal lyrics control of any model here.

Style fidelity you can dial

The prompt-adherence scale (0–20) sets how strictly YuE follows your style prompt — low for loose interpretation, high for exact genre and instrumentation.

Open-source lineage

YuE's open weights keep the model transparent and community-driven — a favorite for creators who want an alternative to closed corporate models.

YuE 2 at a Glance

Price

30 credits

Output

1 song per run

Access

Free tier available

Input limits

Lyrics 3,000 · Style 1,024 · Title 80 chars

Versions & Controls

Custom mode only — lyrics and style are first-class inputs.

Lyrics + style, both required

Custom mode is the only mode: full lyrics (3,000-character limit) and a style prompt are both mandatory inputs. YuE does not auto-write words.

Reference audio (music-to-music)

Attach a track from your uploads or library and YuE generates against it — 40 credits instead of 30 — for style-matched songwriting and covers-with-a-twist.

Prompt adherence, 0–20

One advanced slider controls how literally the style prompt is taken. It stays at the default unless you move it — nudge it when a generation drifts from your brief.

Try YuE 2 With These Prompts

One click drops the prompt straight into the editor above — tweak it, then generate.

pop, female vocals, catchy chorus, bright synth production, 118 BPM

rock, male vocals, gritty guitars, energetic drums, anthemic chorus

ballad, emotional vocals, piano-driven with orchestral strings, slow tempo

About the YuE Music Model

The model

YuE is an open-source song generation project, and YuE 2 marks the current generation served on RaoMusic through hosted inference. Its design centers the writer: lyrics and style are both explicit, first-class inputs, and the model's job is performance, not ghostwriting.

How you drive it

A prompt-adherence scale from 0 to 20 sets how strictly the style prompt is interpreted, and a reference-audio slot turns the same model into a music-to-music tool. There is no simple mode — bring lyrics and a style every time, or draft the words in the AI lyrics generator first.

When to choose YuE 2

Choose it for

  • Lyrics-first: the words you paste are the words it sings, never rewritten.
  • A 0–20 adherence dial and reference-audio mode keep the performance on brief.

The trade-off

No simple mode — lyrics and a style are required every time.

Better fit elsewhere

Compare all six AI music models

How to Generate a Song with YuE

Write the words, tune the style, generate.

  1. 1

    Write the lyrics

    Bring a full lyric sheet and a style prompt — both required.

  2. 2

    Tune and generate

    Optionally set adherence or attach reference audio, then generate one MP3.

  3. 3

    Edit and release

    Extend, cover or mashup the result in the editing suite.

YuE on RaoMusic — FAQs

Yes to both. RaoMusic runs YuE 2 on hosted inference, and every song carries a commercial license at no extra cost — videos, streams, ads and client work included.

No — YuE is lyrics-first by design: your lyrics and your style prompt are both required, and it performs them literally. For AI-written lyrics, use the lyrics generator or MiniMax, then bring the words to YuE.

It controls how strictly YuE follows your style prompt, on a 0–20 scale. Higher values keep the genre and instrumentation closer to your brief; the default suits most songs.

Upload or pick a track from your library and YuE generates a new song against it — same lyrics, guided sound — for 40 credits instead of 30.

Your Lyrics, Performed

Full lyrics control, reference-audio mode and tunable prompt adherence — 30 credits a song, free credits to start.