Give an audio idea a new musical direction. Upload a recording or capture one here, describe the sound you want, and generate music from your reference. Start with a melody, an instrument take, or a musical demo you have permission to use.
Your song
These shared music previews can help you find words for a mood or arrangement. They are listening references from across the site, not demonstrations of what your uploaded audio will produce. Notice whether you prefer a gentle opening, a steady pulse, or a brighter texture before writing your own style description.
Bring one recording and a clear idea of the direction you want to explore. The audio supplies a reference; your description explains the musical treatment. You can prepare both on this page before deciding to generate.
Use Upload audio for a file already on your device, or Record audio to capture a new take with your microphone. Play the preview and make sure you selected the right recording. Choose a passage with an audible musical idea rather than a long introduction of silence. If the take is wrong, remove it and select another before continuing.
Write a style description of 10 to 300 characters. Combine a genre, a mood, and one or two instruments: for example, warm acoustic folk with fingerpicked guitar, soft percussion, and a relaxed feel. Choose details that belong together. A short, coherent direction is easier to evaluate than a list of conflicting genres or a request to reproduce a particular performer.
Sign in when prompted and review the credit amount shown on the generation button. Submit when your audio and description are ready. The result panel follows the task and presents the available music when it finishes. Listen through the new version, then use the result controls to keep a download or try another direction with a revised description.
Audio to Music starts from something you can hear. Pick the kind of reference that best communicates your idea, then give the new version a distinct purpose. These approaches help you choose a useful starting point without needing a finished production.
A keyboard phrase or guitar recording can hold an idea that is difficult to explain in words alone. Select a section where the phrase is easy to hear, and use the description to suggest a different setting around it. You might explore a soft, intimate sound for a reflective piece or a more rhythmic arrangement for an energetic demo. Keep the source recording so you can compare the musical choices afterward.
Make music from audio
When a musical phrase arrives away from your usual setup, use the microphone option to capture a take here. Find a quiet spot, keep a comfortable distance from the microphone, and leave a little space before starting. Listen once before submitting. For a melody made entirely by humming, the dedicated Hum to Song page offers the same recording workflow with guidance focused on that kind of input.
Make music from audio
Use a demo you created as a reference for an alternative musical direction. Decide what you want to explore before writing: a lighter instrumental texture, a slower mood, or a different genre. Express that choice in ordinary language. The result is a newly generated interpretation, so compare the overall feeling and arrangement rather than expecting the original recording to remain untouched underneath added instruments.
Make music from audio
Before generating a second version, write down one thing you want to change. Perhaps the first feels too busy for your sketch, or you want a more spacious sound. Update that part of the style description and listen again. Keep notes alongside downloaded versions so you can remember which direction each one explores. This makes the next creative decision more concrete than simply asking for a better song.
Make music from audio
A few practical choices can make the upload and review process simpler. Check the source before you spend time shaping the style description.
MP3 and WAV are good starting choices. The file picker also accepts M4A, OGG, WebM, and FLAC, but the browser must be able to decode the actual audio inside the file. A filename extension alone does not establish compatibility. If a file cannot be read, export another copy from your recording app and select it again.
Uploaded audio must be between 6 and 360 seconds and no larger than 20 MB. Recording on this page stops at 60 seconds. Prepare a suitable excerpt in your own audio editor if your source is longer; there is no trimming timeline here. Let the selected passage communicate the idea without unrelated takes at either end.
Start with a melody, an instrument performance, or a musical demo. Check for overpowering room noise, accidental handling sounds, and clipping before choosing the take. These are source-selection tips, not automatic repair controls. For a spoken note that describes a song rather than performs one, write the idea in the AI Music Generator instead.
Use your own recordings or material you have permission to submit. Keep a local copy of the source and consider whether it contains private conversations or someone else's performance. Generation uses uploaded audio as a reference; permission to upload and permission to publish a finished track are separate questions to resolve for your project.
Find the right input and understand what this audio-based workflow gives you before submitting a recording.
Choose the recording that best captures your idea, give it a musical direction, and hear a new interpretation. Return to the tool to upload a file or record a take, then make your first style choice.