How to Clone Yourself with AI: Text, Voice and Video Options

1Three ways to clone yourself

An “AI clone of yourself” can copy one of three things: how you write, how you sound or how you look. Some services combine two or three, but each part is built separately and from different material.

A complete “digital you” that thinks, remembers and decides like you does not exist. If you are curious about the bigger promises, our article on what an AI clone is explains the limits plainly.

2Before you start: purpose, consent and security

• Decide what the clone is for. Talking with a version of yourself, leaving something for your family and making marketing videos need different routes.
• Clone only yourself, or someone who clearly agrees. Most voice and video services ask you to confirm consent, and several check it.
• Read where your material goes. Check whether recordings or chats are stored on a server, for how long, and whether they are used to train models.
• Plan how you will label it. Anything your clone produces should be presented as AI, not passed off as you.

3Route 1: a text clone from your messages

A text clone captures the most about your personality, because years of messages hold your vocabulary, your jokes and the people in your life. There are two ways to build one.

The quick way is to paste some of your writing into a general chatbot and ask it to imitate you. It can catch your tone for a few replies, but it does not see your whole history, and pasting private chats into a cloud service has costs of its own. Read whether it is safe to upload your chats to an AI before you try this.

The thorough way is a tool that imports your chat exports, indexes them and learns your style. The main steps look like this:

1. Export your own chatsDownload conversations where you write a lot, for example from WhatsApp, Messenger, Instagram or Discord.
2. Add notesWrite short notes about facts that never appear in chats: where you grew up, people who matter, stories you often tell.
3. Import and confirm who is whoLoad the exports into the tool and tell it which sender is you.
4. Test it honestlyAsk it to answer real messages you received, then compare its replies with what you actually wrote.

Our full walkthrough, how to create an AI clone of yourself, covers each step in detail.

4Route 2: a voice clone

A voice clone is a text-to-speech model tuned to your voice. You type, it speaks in a voice that sounds like yours. Many services offer a fast version from a short sample and a better one from longer recordings.

ElevenLabs is a well-known example. Its documentation recommends about one to two minutes of clean audio for Instant Voice Cloning. For Professional Voice Cloning it suggests at least thirty minutes and ideally two to three hours, followed by a voice verification step in which you read a short text so the service can check the voice is yours. Apple takes a different approach: its Personal Voice feature, built into recent iPhone, iPad and Mac systems, asks you to read a series of sentences and processes the voice on the device, for personal and non-commercial use.

1. Choose a quiet roomSoft furnishings reduce echo. Turn off fans, music and notifications.
2. Use a decent microphoneKeep a steady distance and speak naturally, in the tone you want the clone to use.
3. Record and uploadFollow the service’s length and format guidance, and confirm that the voice is your own.
4. Complete verificationMany services ask you to read a phrase live so they can match it to your samples.
5. Protect the resultKeep the voice private unless you have a clear reason to share it.

A voice clone on its own has no personality; it reads whatever text it is given. Our comparison of voice clones and text clones explains what each can preserve.

5Route 3: a video avatar

A video avatar is an animated version of you that presents a script on screen, often with a cloned voice. Two widely used services show the typical process. HeyGen says a custom avatar starts from a short clip of yourself (it advertises about fifteen seconds) and an on-camera consent code that verifies your identity. Synthesia’s personal avatars can be recorded with a webcam or phone, require a recorded consent video, and include a clone of your voice.

1. Set up light and framingFace the camera in even light, with a plain background and your face clearly visible.
2. Record your footageSpeak and gesture naturally for the length the service asks for.
3. Record the consent stepRead the consent statement or code on camera, as the same person who appears in the footage.
4. Write a script and generateType what the avatar should say, review the result and correct anything that looks or sounds wrong.

Video avatars are built for presenting. They know nothing about you beyond the script.

6Security: voice fraud and deepfakes

Voices and faces are used to prove identity, which makes these clones attractive to criminals. The US Federal Trade Commission warns that a scammer may need only a short audio clip, perhaps taken from a video posted online, to fake a relative’s voice in a “family emergency” call.

• Agree a private code word with close family, and call back on a number you know before acting on an urgent request for money.
• Avoid posting long, clean recordings of your voice publicly if you do not need to.
• Label AI-generated audio and video clearly. In the EU, the AI Act requires deepfakes to be disclosed as artificially generated or manipulated.

Laws on voice, likeness and AI differ by country; this is general information, not legal advice. Our guide to the legal and ethical side of AI clones covers consent in more depth, and the ethics checklist before cloning someone helps if anyone else is involved.

7What to do with your clone

Once you have a clone, the useful part is how you use it. A text clone can become a mirror for reflection: see how to talk to an AI version of yourself, try texting your past self, or use AI to write a letter from your future self. Whatever you build, it can be wrong, it can flatter you and it is not a substitute for people or professional support.

8Where Memory Clone fits

Memory Clone covers the text route only. It is a Windows desktop app, currently in alpha, that builds a private AI simulation from chats you export yourself (WhatsApp, Messenger, Instagram and Discord) plus your notes. It learns your style separately for each relationship and keeps imported originals and the memory index on your PC.

• Cloud processing is off until you switch it on for a clone; then a bounded selection of excerpts is sent to write replies. Your content is not used to train models.
• It does not use voice notes, photos or video today, and it does not clone voices. Audio and video support is planned, and voice cloning is a research item.
• You must be 18 or older, and cloning another living person requires their permission under our Acceptable Use Policy.