Clone Your Own Voice Offline on iPhone

Make and use your voice

Open Voices and create a voice. Choose to record a reference sample, or import audio or a video from Photos or Files.
Choose a clear sample. The clone uses up to the first minute of the sample. If your file is longer, you can choose to use its first minute.
Name and save it. Saving a personal voice requires Pro. The first use of a neural model may include a download or preparation step.
Make a clip. Record, import, or start with typed text; select your saved voice, generate the result, then save or export it.

Use only your own voice or one you have permission to clone. VocalShift is for recorded or generated audio, not live calls.

A clean, single-speaker sample is a better starting point than noisy or mixed audio. Importing a clone sample does not separate vocals from background music.
Voice cloning changes the voice timbre; it does not make a live phone or FaceTime call use your clone.
VocalShift needs iOS 18 or later. A network connection is needed for the initial model download and may be needed for purchases.

Can I import a sample instead of recording one?: Yes. You can select an audio file or video from Photos or Files. Only the first minute is used for the personal voice.
Can I clone another person's voice?: Only if that person has given you permission. Do not use a clone to impersonate or deceive someone.
Do I need to train a model on a computer?: No. VocalShift creates the personal voice on the iPhone from the reference sample.