What is a custom avatar or Digital Twin? Learn how it works, what it costs, and how to create yours from a short recording. Answers to the top FAQs inside.
TL;DR: The fastest path is recording about 30 seconds of yourself in HeyGen and letting it build a Digital Twin on Avatar V, because one consented recording becomes a presenter that keeps your face, gestures, and cloned voice across new scripts and 177+ languages. Most twins are ready in 10 to 20 minutes.
- Synthesia's personal avatars are the better fit for teams already producing on its plans, which include 3 on Starter and 5 on Creator.
- Voice-only cloning in ElevenLabs is still the better call for podcasts and narration where no face is needed.
- Batch-filming real videos yourself wins when you enjoy being on camera and only need a few weeks of content.
It's Sunday night. An onboarding module, three LinkedIn clips, and a client explainer all need your face by Wednesday, and you're on take nine of the same opening line.
A digital twin avatar is the fix: a reusable video version of you, trained on a short recording, that delivers any script you type. This guide explains what a custom avatar is, how to create one in five steps, what it costs, and the rules nobody mentions until after you record.
The platform behind it holds a 4.8/5 G2 rating from 1,901 reviews. Here's the full walkthrough.
How to Create a Digital Twin with HeyGen
1. Start the Clone a Real Person flow
Log in, open Avatars, click New Avatar, and choose Clone a Real Person. You can record live through your webcam or phone, or import footage you already have.
The Free plan includes one Digital Twin slot, and each slot holds up to 500 looks of the same person. Getting to the recording screen takes about 2 minutes.
2. Record the footage your twin will learn from
Film one continuous take at 1080p and 30fps or higher, with your face visible and no cuts. The platform's own guidance asks for about 30 seconds to camera, and longer clean footage helps facial range.
Be more expressive than feels natural, because your twin's gestures and expressions come straight from this clip. If you under-act here, the AI video avatar you get back will look stiff. Budget 10 minutes, including setup.
3. Record your consent clip
Next, read the on-screen consent code to camera. The person in the consent clip must be the same person as in the footage, and screen recordings or AI-generated clips fail verification.
This step takes about 2 minutes. It's also why nobody can build a twin of you from a video they found online: the consent has to come from your own face and voice.
4. Clone your voice
Each video avatar includes voice cloning, and a cloned voice gives a twin the closest match. Record a clean sample in a quiet room, with natural energy rather than a flat read.
Avatar V weighs audio first when it decides how your twin moves, so expressive audio produces livelier gestures. Once AI voice cloning finishes, your twin uses that voice in its preview automatically. Allow about 5 minutes.
5. Generate your first video and check it
Wait for training: most twins are ready in 10 to 20 minutes. Paste a 30-second test script, confirm Avatar V is selected, and click Generate.
Don't judge the mouth in the editor preview, because lip sync only renders on submission. Check the finished render for eye line, gesture variety, and pronunciation of names. Avatar V uses 48 credits per minute, so a short test protects your monthly balance.
What You Need Before You Start
- A camera at 1080p. A recent phone or a good webcam works. Prop it at eye level so the twin looks at viewers, not past them.
- A quiet, evenly lit room. Face a window or a soft light, and turn off fans. Good enough means no harsh shadow across half your face.
- A plain outfit. Skip large logos, text, and busy patterns. Glasses and jewelry are allowed but can cause small artifacts.
- Something to talk about. Explain your work or answer a customer question in your own words. A natural ramble beats a memorized script.
- The real person's time. For a twin of a CEO or colleague, that person must record both the footage and the consent clip themselves.
Common Mistakes to Avoid
Under-acting in the source footage
Your twin copies the hand movements and expressions in the training clip. If you sat still and spoke in a monotone, the twin will do the same in every video. Record with more energy than feels comfortable, and include natural hand movement below the chest.
Recording with cuts or a moving camera
Uploads get rejected for multiple scenes, too much head movement, blur, or poor lighting. Walking while filming on a phone is the most common cause. Mount the camera, stay in one spot, and record a single continuous take.
Leaving dead air at the end of your script
A TechRadar reviewer found his twin silently vamping for the camera for half the video because of blank space at the end of the script. Trim trailing spaces and line breaks before you click generate, and preview the timeline length.
Building a twin of someone without them
Consent fails when the consent clip doesn't match the training footage. Footage of a colleague, a client, or a public figure won't pass without that person recording their own consent. Plan the session with them in the room.
The Rules You Only Learn After You Record
None of the ranking guides explain what you can't change once your twin exists. These rules come from the platform's own help documentation.
- One redo per billing cycle. Re-recording means deleting the avatar from its slot and uploading new consent and footage, and you can do that once a month.
- One person per slot. A slot's 500 looks all belong to the same individual and the same consent video. A second person needs a second slot.
- Gestures are fixed by your footage. Prompts can't add hand movements your training clip never showed, so the recording session matters more than any setting.
- Slot counts depend on your plan and account. The help center lists 5 slots on Creator and Pro and 10 on Business, while the pricing table shows 1+ and 5+, and both note that included slots vary by region and account age. Extra slots cost $29 per month or $300 per year.
What Results to Expect
Educator Anton Voroniuk built his course production around avatars instead of filming every lesson. He saves 15.5 hours per week and has taught more than 1 million students. You can read Anton Voroniuk's story for the workflow behind those numbers.
Reid Hoffman's digital twin, built on HeyGen's LiveAvatar for real-time conversation, has reached more than 50 million impressions. It shows how far a twin can travel once the recording is done.
A G2 reviewer described the core benefit in one line: film once, then feed scripts for future videos. That reviewer's team used it for regular founder videos while the founder was rarely available.
Other Ways to Put Yourself on Video
Synthesia personal avatars
Synthesia is an avatar platform widely used for corporate training. Its pricing page lists 3 personal avatars on Starter and 5 on Creator.
- Pros: personal avatars included in paid plans; a slide-style editor familiar to PowerPoint users; one-click video translation; an established enterprise customer base.
- Cons: monthly video minutes cap output, with Starter covering 10 minutes a month; the Free plan includes no personal avatars, so you can't test your own likeness before paying.
Voice-only cloning in ElevenLabs
ElevenLabs is a voice platform that clones your voice for narration, voiceovers, and dubbing.
- Pros: high-fidelity voice output; fast generation; strong language coverage; a natural fit for podcasts and audiobooks.
- Cons: no face, no lip sync, and no presenter, so video needs a second tool; you still have to source visuals for every script.
Batch-filming yourself
Block a morning, film a dozen scripts back to back, and release them over several weeks.
- Pros: the real you, with no AI disclosure questions; full control over delivery; high quality with a good setup; no subscription required.
- Cons: every script change means re-filming, which breaks down for training content that updates quarterly; output is capped by your calendar and energy.
Frequently Asked Questions
Is a Digital Twin the same as a deepfake?
No. A deepfake uses someone's likeness without permission, while a Digital Twin requires a live consent recording from the person themselves. Your custom avatar is tied to your workspace, HeyGen doesn't make it available to other users, and content on every account goes through moderation.
Do I have to tell viewers it's my avatar?
Often, yes. YouTube asks creators to disclose realistic synthetic content at upload, and it began auto-labeling photorealistic AI video in 2026, while TikTok has an AI-generated content toggle for realistic avatars and cloned voices. Disclosing up front also protects trust with your audience, which matters more than the label itself.
Can my twin speak languages I don't speak?
Yes. Your twin can deliver scripts in 177+ languages and dialects on paid plans, with lip movement matched to each language and your cloned voice carried across. For existing videos, the video translator re-syncs your twin's lips to the new audio, and a native speaker should still review the script.
Can teenagers create a Digital Twin?
Yes, as long as the person can provide consent themselves. Avatars of users of any age go through the same consent check and content moderation as adult avatars. For school or youth programs, a parent or guardian should review where the videos will appear.
What if I don't want to record video at all?
Start with a photo instead. One front-facing photo becomes a talking avatar through Avatar IV, with natural expression and lip sync. An AI photo avatar won't copy your real gestures the way a video-trained twin does, but it needs no camera.
How many Digital Twins can I have?
The Free plan includes one slot, and paid plans include more, though exact counts vary by plan, region, and account age. Each slot holds one person with up to 500 looks. Extra slots cost $29 per month or $300 per year, added from the pricing screen.
How much video does the Creator plan cover with a twin?
About 12 minutes of Avatar V video a month. Creator includes 600 credits for $29 per month, and Avatar V uses 48 credits per minute. Switching drafts to Avatar IV, at 16 to 31 credits per minute, stretches the same credits for tests before a final Avatar V render.
Conclusion
Creating a digital twin avatar takes one expressive recording, a consent clip, and a voice sample, and most twins are ready within 20 minutes. The Free plan includes one Digital Twin slot to try it, and Creator starts at $29 per month, or $24 billed annually.
Greetings! My name is Ayesha Shaheryar. My words have helped millions over the past two years. As a HeyGen expert and a writer, I am here to introduce tips and tricks to edit your next video in no time.
