welcome to Musicfy! We are super excited to see the amazing music you will be able to produce with your custom AI voice clone.

this voice clone is completely private and only you have access to it.

Getting Started

• dry acapellas work best (no reverb, echo or other vocal processing)
• do not have any background noises, instrumental, bass, or auto tune
• supported File Types: MP3, WAV, M4R
• if possible remove 1 second+ silences between words/lines
• try and use vocals where it's just one person singing/no layering

Bad Training Vocals

bad_stems.wav

❌ contains background noise

❌ contains background harmonies

Good Training Vocals

good_stems.wav

✅ no reverb or echo

✅ no background noise

Tips & Tricks

• what you give us is what your clone can do
• if you give us rapping your clone can rap. If you do singing in a certain style your clone will do singing in that style.
• a decent data set will be 3-5 minutes of vocals. A good data set is around 10 minutes of vocals.
• the more minutes you can give us the more accurate the model will be but the longer it will take to train.
• cover a range of tones and notes (i.e. sing & rap, cover your entire octave range)

Enjoy your Clone

<aside> 🎤 once your stems are uploaded, please expect a ~2 hour turnaround time for your model to be on the Musicfy platform.

</aside>

Publishing A Community Model