AI Music Generator
Describe a mood and get an instrumental MP3.
- Free
- No sign-up
- Runs on GPU
Files are deleted automatically after 24 hours. We do not log your content or use it for AI training.
Type a short description of the music you want, or start from one of the genre presets, choose a length between 10 and 60 seconds, and the generator composes an instrumental track for you. The result plays in your browser and downloads as an MP3.
The music is made on our own GPU server with the open-source ACE-Step 1.5 model (MIT licence). It has no singing: every track is instrumental. A track usually takes about a minute to appear, and a little longer when the model has to start up first.
How to use AI Music Generator
- Describe the music in English, or press a preset such as calm ambient or lo-fi hip hop.
- Move the length slider (10 to 60 seconds).
- Press Generate music and wait for the progress bar to finish.
- Listen, then download the MP3. Use the same seed number in the API to repeat a result.
Features
Calm ambient, pop electronic, cinematic orchestral, lo-fi hip hop and acoustic folk, each a good starting caption you can edit.
No lyrics and no vocals, so it works as background music.
From 10 to 60 seconds per track, as an MP3 file.
Frequently asked questions
Which languages can I describe the music in?
English works best. The description is a style caption such as "warm piano, slow tempo", not lyrics.
Can I use the music commercially?
The model is open source under the MIT licence. You are responsible for your description and for how you use the track. Results can be similar to existing music by chance, so check before you publish.
Is my description private?
It is used only to compose the track. We do not log it or train models on it, and the file is deleted automatically after 24 hours, or at once with the Delete now button.
Why does it take so long?
The music model is large, so it is loaded only when needed and shares one graphics card with the other tools. Expect about a minute per track.
