Make your iPhone speak in a Personal Voice you create

An iPhone can speak a new sentence in a synthetic voice created from its owner’s recordings. Personal Voice makes that voice, while Live Speech supplies the words the person types. Apple presents the combination as a communication aid for people whose speech is affected or who are at risk of losing it.[1]

Apple’s current accessibility page describes recording ten randomised phrases and generating a voice in under a minute. It also offers short, three-word phrases for people who find full sentences difficult. Those are Apple’s documented capabilities and timing claim. This walkthrough follows its instructions; it is not a timed test of voice creation.[1]

Record the voice on your phone

This guide follows Apple’s current iOS 27 instructions, which list iPhone 15 Pro models, iPhone 16 models and later. The listed languages are US English, mainland-China Mandarin and Mexican Spanish. Older software has separate documentation, so check the guide for the version you use.[2]

Open Settings → Accessibility → Personal Voice → Create a Personal Voice, then follow the recording prompts. A passcode must be set before the creation option appears. You can pause a session and resume the recording from that voice’s entry later.[2]

How a recording becomes new speech

The AI has to produce words that were never spoken in the recordings. Apple’s December 2023 technical account explains the original system as a sequence: convert written words into speech sounds, predict their acoustic representation, then generate an audio waveform. Personalisation adapts a previously trained speech model to characteristics of the recorded voice.[5]

That first generation used 150 recorded sentences and overnight processing. Its architecture explains the task, but does not establish which models run inside today’s faster version. The same research also describes why recording quality matters: background noise can make vocal characteristics harder to learn, and its system used speech enhancement to clean recordings. A quiet recording space remains a sensible starting point.[5]

Choose the voice and start a conversation

After creation, open Settings → Accessibility → Live Speech and turn it on. Under Voices, select the language and the voice you created. Apple’s voice-selection support page describes this control; the current iPhone guide supplies the conversation and phrase steps below.[3][4]

Triple-click the side button, choosing Live Speech if several accessibility shortcuts appear. Tap its window, type or paste your sentence and tap Speak. The current guide says FaceTime participants hear the output at their end; outside a call it plays through the iPhone speaker.[3]

For repeated wording, return to Live Speech settings and open Phrases. Add a category, open it, add a phrase and save it. During use, open the Phrases window, select the category and tap the phrase you need.[3]

Try one saved sentence, then compose a reply that fits what someone has just said. Listen for intelligibility, a comfortable pace and whether the voice feels recognisable to you. This rehearsal checks the complete conversation task: choosing the words, finding the controls and being heard. Change the examples to language you actually use.

Choose where the voice is available

Personal Voice settings include Share Across Devices for iCloud-linked use, alongside a switch allowing compatible communication apps to request access. Apple restricts creation to your own voice for personal, non-commercial use. These controls are worth deciding during setup, while the phone is already open to the relevant settings.[2]

A useful first outcome is modest: a recognisable voice ready to say a few familiar things, with room to write an unexpected answer. The words remain the speaker’s choice. Preparing that combination is the practical purpose of the recording and setup.

Report an error or suggest a correction