How to Make an AI Video of Yourself, in Your Own Face and Voice

There are four routes: a digital twin built from your footage, a quicker photo avatar, Persona Video inside your Persona, or plain video you film yourself. Pick by who will watch and how many videos you need.

• Footage-based digital twin. You film a short clip, the tool builds an avatar that moves like you, and you type a script for each video. HeyGen asks for about 30 seconds to camera and says longer clean footage helps facial range (HeyGen, 2026). This route looks the most convincing and suits anyone who publishes every week.

• Photo or selfie avatar. Captions says it can generate your clone from a single selfie, and HeyGen's free plan includes up to three photo avatars (Captions and HeyGen, 2026). A still photo carries far less of how your face moves than real footage does.

• Persona Video inside Afterlife.ai™. You build your Persona from your own memories, stories and voice, and Persona Video makes video messages of you from one 30-second recording. The people you choose can also talk with your Persona.

• Film it yourself. A phone video costs nothing, needs no consent step and looks exactly like you, but every new message means another take.

Use HeyGen for a free test or marketing at volume, Synthesia for company training, Captions for social clips, and Afterlife.ai™ Persona Video for messages to your family. Prices and limits below come from each company's own pricing page (HeyGen, Synthesia and Captions, 2026).

• Best free option: HeyGen, with one custom avatar, one voice clone and three one-minute videos a month.

• Best for company training: Synthesia, with personal avatars from the $29 Starter plan.

• Best for social clips: Captions Max at $24.99 a month, with digital twins and AI Voice Clone.

• Best for family video messages: Afterlife.ai™ Persona Video, in your own cloned voice, inside a Persona your family can talk with and Executor Lock™ protects.

• Best for one video this week: your phone, because one message does not need a subscription.

All three studios meter usage in credits, which hides the real cost per video; HeyGen says the credits a video uses depend on the AI model, the length and the complexity of the generation (HeyGen, 2026). For more apps, see our best AI avatar app comparison and our best AI clone apps guide.

One short AI video of yourself can cost $0 on HeyGen's free plan or your own phone. A weekly habit runs about $216 to $348 a year on the studio tools (prices as of October 2026).

Example: you want to send your granddaughter a one-minute video every week for a year, 52 videos in all. HeyGen's free plan stops at three a month, so you need Creator. Synthesia needs Starter for a personal avatar, and its minutes (10 free a month plus about 12 more) easily cover four or five one-minute videos. Captions prices in credits rather than minutes, so test one month before you pay for a year.

On Afterlife.ai™, the free build and your first Persona Video cost nothing, with no card. Legacy is $14.99 a month and Eternal is $29.99 a month, and the pricing page lists what each plan includes.

Pick the tool, film one clean 30-second clip, record any consent statement, clone your voice, then write a short script and generate. The clip takes most of the effort, because every video the tool makes copies it.

1. Decide who the video is for. A sales explainer and a message to your son need different tools.

2. Pick your tool from the table above and read its consent rules before you film.

3. Film your training clip: about 30 seconds to camera in one continuous take, using the checklist below.

4. Record the consent statement if asked. HeyGen wants a clip under 30 seconds, read exactly as written by the person in the footage (HeyGen Help Center, 2026).

5. Clone your voice, because a twin with a stock voice looks like you and sounds like someone else; our voice cloning guide covers how to record clean samples.

6. Write a short script the way you talk, and read it out loud once before you paste it in.

7. Generate, then watch twice: once with sound and once on mute, which shows mouth and gesture errors.

8. Keep the script and the clip together, in case you change tools later.

Film one continuous take at eye level, in a quiet room with even light, and be more expressive than feels natural. HeyGen warns that if you sit still and speak in a monotone, the twin does the same in every video (HeyGen, 2026).

• Camera: a recent phone or good webcam at 1080p and 30 frames per second or higher, on a stand.

• Height: the lens at eye level, so you look at the viewer rather than past them.

• Light: face a window or a soft lamp, with no harsh shadow across half your face.

• Sound: a quiet room, with fans, air conditioning and music off.

• Framing: your whole face, with no cuts, no walking and no turning away from the lens.

• Clothes: a plain top without big logos, text or busy patterns; glasses and jewelry are allowed but can cause small artifacts.

• Hands: natural gestures, because HeyGen says prompts cannot add hand movements your clip never showed.

• Energy: a bit more expression than feels normal, with real smiles and real pauses.

Delivery matters more than content in this clip. Tell a story you know by heart, such as how you got your first job, so you talk at your natural pace and your face moves the way it does at dinner. Do two or three takes and keep the one where you forgot the camera.

Say something only you could say, to one named person, in under a minute. The best first videos are specific: a story, advice with the reason behind it, or a thank-you for one moment.

• The day they were born, from your side: where you were and what you thought the first time you held them.

• How you met their grandmother, including the part you usually skip.

• Your best and worst money decisions, with the numbers.

• What you wish someone had told you before your first job.

• The family recipe, with the step everyone gets wrong.

• Five things you are proud of them for, with a reason for each one.

• Where the family comes from: the town, the move and why your parents made it.

• The running joke only your family understands, and how it started.

Open with the person's name and the reason for the message, tell one story, then say what you want them to take from it. For help writing scripts in your own words, read our guide to an AI that sounds like you.

Build your Persona free, create your voice, record one 30-second clip, and make your first Persona Video free. Everything runs in the iPhone and Android apps or at app.afterlife.ai, on one account.

1. Start your free build in the app or at app.afterlife.ai: 25 memories to build your Persona, up to 5 Trusted Contacts and Executor Lock™ setup, no card, and your free build never expires.

2. Add your memories: the stories, people and opinions that make you who you are, in your own words.

3. Create your voice. You record your own voice, and your Persona speaks in your own cloned voice; creating your voice is free, and so is your first listen.

4. Record your 30-second clip with the filming checklist above.

5. Make your first Persona Video, free, from one of the ideas above.

6. Add Trusted Contacts and name your executor to set up Executor Lock™.

7. Pick a plan when you want more: Legacy at $14.99 a month, Eternal at $29.99 a month, or Afterlife Pro at $59.99 a month or $599 a year, which adds Afterlife.ai™ Twin. See the pricing page.

For family messages, a Persona beats a video studio. Your family gets the message in your face and your own voice, and they can talk with your Persona in your own words, in 42 languages. A studio avatar reads whatever script is typed into the account, while your Persona is built from your own memories and stays under your control.

Executor Lock™, our patent-pending governance layer, is how that control works. You name your executor and Trusted Contacts while you are alive, and at verified death your Persona is locked as a perfect snapshot. Nobody can edit, retrain or impersonate you after that, so your family always meets the Persona you approved, and the things you leave for people are released after Executor Lock™ activates. Idy® has 78 patents pending, including two international PCT applications filed in August 2026, and the Executor Lock™ page has the details.

Most fake-looking AI videos come from the source clip, not the software: stiff delivery, a moving camera, harsh light or a noisy room. Fix the clip and you remove the most common causes.

HeyGen's 2026 guidance says walking while filming on a phone is the most common fault, and uploads get rejected for multiple scenes, too much head movement, blur or poor lighting. HeyGen also says its Avatar V model weighs your audio first when deciding how the twin moves, so expressive audio produces livelier gestures (HeyGen, 2026). The other giveaways are the voice and the script: a stock voice coming out of your face is the first thing your family will notice, so clone your own voice before you generate anything. Then write with your own phrases and the names you actually use, and cut any line you would never say out loud.

Yes. No US law stops you from making an AI video of yourself, and the fine print is the tool's own terms. Making one of someone else without their consent can break state and federal law, and those rules tightened between 2024 and 2026.

• TAKE IT DOWN Act (federal). Public Law 119-12, signed on May 19, 2025, makes publishing non-consensual intimate images a federal crime, including AI-made "digital forgeries," and requires covered platforms to remove such images within 48 hours of a valid request. Platforms had a year to build that process, so the duty applies from May 19, 2026 (US Government Publishing Office, 2025).

• ELVIS Act (Tennessee). Governor Bill Lee signed the Ensuring Likeness Voice and Image Security Act on March 21, 2024, effective July 1, 2024. It adds voice, including simulated voices, to the state's right of publicity and reaches anyone who makes available a tool whose primary purpose is producing a person's voice or likeness without authorization (Davis Wright Tremaine, 2024).

• NO FAKES Act (federal, proposed). The Senate Judiciary Committee advanced the NO FAKES Act of 2026 (S. 4591) by unanimous voice vote on June 18, 2026. It would create a federal right to control your voice and visual likeness in AI digital replicas, with notice-and-takedown for platforms, and after that vote it still needed the full Senate and the House (Holland & Knight, 2026).

• State right of publicity. Many states protect a person's name and likeness against unauthorized commercial use, and our digital replica laws by state table covers each one.

The better tools enforce consent themselves. HeyGen requires the person in the avatar footage to record their own consent video (HeyGen Help Center, 2026), and Synthesia says no avatar is created without informed consent through a biometric consent process, with permissions people can revoke at any time (Synthesia, 2026). Captions' clone-yourself page describes a clone from a single selfie and no consent step, so the burden sits with you.

Afterlife.ai™ is strict on purpose. A Persona is built by the living person from their own memories and voice, Afterlife.ai™ never recreates a person who did not consent while alive, and under 18s cannot build a Persona. A grown child can help a living parent build the parent's own Persona, but the parent owns that Persona and makes every decision.

• HeyGen: Pricing Plans for Creators and Marketers

• HeyGen Help Center: Consent video requirements for digital twins

• HeyGen: How to Create a Digital Twin Avatar: What It Is and FAQ

• Synthesia: Pricing

• Synthesia Help Center: Avatar Policy Foundation

• Captions: Pricing and Subscription Plans

• Captions: Clone yourself with AI

• US Government Publishing Office: Public Law 119-12, the TAKE IT DOWN Act

• Davis Wright Tremaine: Tennessee, All Shook Up Over AI-Generated Voice Replicas, Passes ELVIS Act

• Holland & Knight: Senate Committee Advances Bill to Protect Name, Image, Likeness and Voice Against Unauthorized AI Use