Upload the song you want to re-sing, choose an AI voice, select the original vocal gender, and start generating. For better results, use audio with clear vocals and minimal background noise.

Songs with clear vocals, steady volume, and less background noise usually work best. If the audio has heavy reverb, obvious noise, or vocals buried under the instrumental, the result may sound less natural.

The original vocal gender helps the system adjust pitch more accurately. Choosing the right option can reduce unnatural pitch shifts and make the AI cover sound smoother.

MelodyCraft supports MP3, WAV, M4A, AAC, and FLAC. MP3 is recommended because it uploads faster and works well for most AI cover tasks.

We recommend uploading audio under 8 minutes and keeping the file size under 100MB. Shorter and cleaner audio usually generates faster and produces more stable results.

Yes. You can preview available public voice models before choosing one for your AI cover, so you can pick a voice that fits the song.

Yes. After the AI cover is generated, you can listen to the result and download the audio if this feature is available in your account.

Commercial use depends on the rights of the original song, the selected voice model, and the laws in your region. Make sure you have the necessary permissions before publishing, monetizing, or distributing AI cover songs.