MusicGPT studio

Make the first version while the idea is clear.

Describe the sound, set the master output, then create and save an original track with MusicGPT.

Rework an existing song with AI cover

Session ready

01 / COMPOSESong direction
Add the lyrics you want the vocalist to sing.Need a lyric draft? Open the AI Lyrics Generator
4 credits per song

How the Studio fits the workflow

Make a brief, choose the delivery, then review the result.

The Studio is the place to turn a written music direction into playable audio. Start by describing the scene, listener, mood, pace, instruments, and vocal choice. You can use your own lyrics, ask MusicGPT to develop lyrics from the brief, or switch to instrumental mode when the track needs to sit under a video, podcast, or other spoken content.

The output settings are part of the decision. MP3 is convenient for a first listen and sharing, WAV is useful when you plan to edit or mix, and PCM is for workflows that specifically need raw audio. One complete song uses 4 credits; check the balance and the current MusicGPT plans before you generate at volume.

01

Write the listening brief

Give the track a job and a point of view. A scene and emotional movement are more useful than a pile of disconnected adjectives.

02

Choose voice or instrumental

Use your lyrics, a generated lyric direction, or no vocal when the music must leave room for narration and visuals.

03

Judge the first version

Listen for purpose, arrangement, delivery format, and the one detail you want to protect in the next revision.

Before and after Generate

The Studio is a decision loop, not a one-click music shortcut.

Before you generate, write the listening brief in the same order a collaborator would need it: where the music will be used, who will hear it, what should happen emotionally, which two or three sounds matter, and whether the result needs a vocal or instrumental role. A scene such as a late train home, a night-time greenhouse, or a product reveal gives the model a usable world. A long string of praise words does not necessarily give the track a job.

Then choose the lyric path deliberately. Use your own lyrics when the words and point of view are already established. Ask MusicGPT to develop lyrics when the story is clear but the draft is not. Choose instrumental mode when the audio must leave room for narration, dialogue, captions, or a visual edit. The choice changes how you should judge the result: a song needs lyrical and vocal fit, while a cue needs timing, density, and space.

Output settings are part of the handoff. MP3 is convenient for a first listen and sharing, WAV is useful when the track will be edited or mixed, and PCM is for workflows that specifically need raw audio. One complete song uses 4 credits in the current Studio workflow. Check the visible balance and the current plan details before generating at volume, and save the prompt and selected format with the project.

After the result is ready, listen for the purpose before judging polish. Identify what should stay, what is getting in the way, and what single change would make the next version more useful. If the output will be public or commercial, review the usage-rights guide, supplied material, platform rules, and client requirements separately from the creative revision.

A practical Studio brief can follow this order: destination, listener, scene, emotional movement, sound anchors, vocal or instrumental role, duration, and delivery constraint. This sequence helps the generator understand what the music is for before it has to choose decorative details. If one field does not matter to the project, leave it out rather than filling the prompt with generic praise.

When you are comparing versions, keep the project goal and the output setting stable. A WAV handoff for editing, an MP3 for a quick share, and a PCM workflow may require different next steps even when the musical idea is the same. Save the chosen format with the prompt so a collaborator knows which version they heard.

The Studio works best as a repeatable review loop: brief the idea, generate one version, name the strongest decision, change one weak detail, and listen again. That process creates useful evidence for the next prompt and helps a team distinguish a creative preference from a technical delivery problem.

When a prompt feels too vague, add one concrete production fact instead of five more genre words. Name the first visual, the kind of voice that must remain audible, the instrument that should lead, the part of the arrangement that should stay restrained, or the exact ending an editor needs. Concrete facts make the next result easier to compare.

01

Brief

Describe the listener, scene, emotion, sound palette, voice, and constraints.

02

Deliver

Choose lyrics or instrumental mode, then select the format and output controls your handoff needs.

03

Review

Keep the strongest idea, change one weak detail, and confirm rights before release.