YuE2 AI Music Generator

Start by signing in and entering a song title and music style. In vocal mode, add structured lyrics with clear line breaks and supported labels such as verse and chorus; instrumental mode lets you leave the lyrics field empty. Use the style field for genre, mood, energy, tempo, vocal character, or instrument ideas. Lyrics provide the words and shape, while the style prompt supplies the sonic direction.

The current YuE2 page uses 20 credits for a generation and returns one track. After you submit the brief, wait for the status to finish before listening in the browser or downloading the result. The inputs guide the interpretation, but they do not expose a score editor or guarantee a particular melody, voice, arrangement, or mix. In vocal mode, the intended result is one complete song with vocals and accompaniment.

A useful prompt gives each field a clear job instead of repeating the same idea everywhere. Keep the lyrics focused on lines and sections, then reserve the style field for sound. A compact direction such as “mid-tempo indie pop, intimate lead vocal, clean electric guitar, and a chorus that gradually opens up” gives the generator a workable arc. You can also name a general energy level or instrument contrast when those choices matter to the sketch.

Describe relationships and movement rather than asking for impossible precision. “Restrained verses with a wider chorus” is a practical creative cue; “copy this exact singer” or “make every note perfect” is not a control the form promises. Keep a short note about the choices you made so the next attempt has a deliberate change. Use lyrics and references you are entitled to use, and treat the prompt as direction for a draft rather than a substitute for your own musical judgment.

The lyrics-plus-style workflow is useful for an original lyric demo, a style experiment, an arrangement reference, or an early songwriting sketch. A generated track lets you hear how the words sit in a melody and whether the vocal delivery supports the mood you described. It can reveal a promising phrase, an unexpected rhythmic idea, or a section that needs more work without requiring you to decide the final production immediately.

Listen with one question at a time: are the lyrics understandable, does the song shape make sense, and does the accompaniment support the vocal? If the answer is no, adjust one part of the brief instead of changing every field at once. You can then download a version that helps you compare ideas and continue the arrangement, editing, or recording in your own production workflow. The first result is a creative reference, not a promise that the song is release-ready.

Symbolic planning describes the internal stage where musical relationships are organized before audio is rendered. In practical terms, it concerns how melody, chords, rhythm, and the larger song structure can fit together around the lyric brief. Thinking of this as a planning layer helps explain why a style prompt can influence the character of a result while still leaving room for interpretation in the finished audio.

This browser workflow does not expose that plan as an editable notation or piano-roll document. You cannot directly move a planned note or rewrite a chord inside the YuE2 page. If the musical shape is not right, use clearer structural or stylistic direction in a later brief, then make detailed timing, harmony, or mix decisions after downloading the audio in a tool suited to that work.

When processing finishes, use the result panel as a listening checkpoint. Check the opening, the transitions between sections, the clarity of the words, and the balance between vocals and accompaniment. Headphones can reveal details that small speakers hide, while a second playback system can show whether the overall idea still holds outside your desk setup. These checks help you decide whether to keep the draft, revise the brief, or continue working elsewhere.

The page provides playback and a download for the available result, and its history view records generations according to the account and history features shown in the app. This model page is intentionally limited to one track per request, so keep a local copy when a version matters to you. Before sharing or releasing anything, review the audio and confirm that your intended use is covered by the relevant rights and terms.

This page is centered on the YuE2 lyrics-and-style workflow. It suits a project that already has words, a rough song shape, or a clear vocal and accompaniment direction. The compact model-specific workspace keeps the model choice visible and makes one generated result the unit you review. That focus can be helpful when you want to learn from one interpretation instead of comparing a larger set of variations.

For a broader set of music-generation controls, compare this page with the general AI music generator. Read the fields and output description on that page before switching workflows, because a general brief-first process and a lyrics-first process ask different questions of you. Whichever route you choose, keep the model label, prompt, and downloaded audio together so you know which brief produced each draft.

This page provides an independent interface and is not the official YuE website. For the project's current code, model information, weights, attribution, and license terms, read the YuE source, YuE2 model card, and model license. Those references are the right place to check details that may change as the project develops.

Code and model-weight licenses can differ, and the service terms for this website are a separate consideration. Before commercial use, confirm that you have the necessary rights to the lyrics, prompts, and any source material, then review the applicable service terms and model licenses for the output you plan to use. Do not treat a general “open source” label as automatic commercial clearance.