Songs from short written concepts
Start with a mood, scene, topic, or story instead of complete lyrics. The text input supplies the creative direction for the generated track.
Text to Song workflow
MusGen Text to Song AI turns a written idea, mood, scene, or story into two song drafts with music and, when instrumental mode is off, generated lyrics and vocals. Choose the available style and voice options, then compare the results.
Lyrics, titles, and styles must not contain famous names or sensitive words.
Describe a mood or story. MusGen writes lyrics, melody, and vocals into a full song.
Text to Song capabilities
This AI music generator from text creates song drafts from a description rather than requiring finished lyrics.
Start with a mood, scene, topic, or story instead of complete lyrics. The text input supplies the creative direction for the generated track.
When instrumental mode is off, Text mode can develop the description into lyrics and a vocal song rather than returning instrumental audio only.
Use the options available in the form to guide genre and vocal character. They influence the result but do not guarantee an exact composition.
Turn on instrumental mode when the idea should become a music-only track without generated sung lyrics.
3-step text workflow
Enter a clear description, choose the available musical options, and start a 3-credit generation. A successful request returns two drafts for you to compare.
01
Write the mood, subject, scene, or story you want the song to express. Specific creative direction gives the generator more useful context.
02
Select a musical style and voice option, add a title if needed, or turn on instrumental mode when the result should not include sung lyrics.
03
Sign in and start the 3-credit generation. When processing finishes, listen to both returned tracks and revise the description if you want another direction.
Text to Song FAQ
Learn what to enter, what the generator returns, and when to use the separate Lyrics to Song workflow.
Describe the concept, choose vocal or instrumental mode, and generate two tracks to compare.