In partnership with

Hey folks,

Google just released Gemini Omni for video and avatar generation so I made an AI version of myself this weekend, and I'm still not sure how I feel about it.

I recorded a short selfie video on my phone, and a few seconds later Gemini handed back an avatar detailed in a way I didn't expect. Not a smoothed-over version. The wrinkles were there. The way my neck moves when I talk. The shape and definition of my face, down to the parts I'd rather it had skipped.

That's the exciting part and the slightly unnerving part, at the same time. You record yourself once. After that, getting a real-looking video of yourself talking is a sentence of typing.

Let me show you how to record your own AI Avatar with Gemini Omni.

Let's dive in..

Together with Harmonic

Claude is not just a chatbot anymore. Is your security team ready?

Claude.ai is one thing. Claude Cowork with MCP connections, running agentic workflows, taking actions across your data with ungoverned skills? That is a different conversation entirely, and most security teams are not equipped to govern it.

Harmonic Security is built to secure everything Claude offers. Full browser controls for Claude.ai, deep governance over agentic MCP workflows, and real-time visibility into what Claude is doing across your organization. So your CISO can say yes to the tools your business is already demanding.

 
What this actually is
 
Google added an avatar feature to Gemini. You record a short selfie video with sound, and Gemini turns it into a personal avatar tied to your Google account. In any prompt, your avatar shows up as @[your username], and you can ask Gemini to generate video with you in it.
 
Who it's for: anyone who keeps needing themselves on camera and keeps not having time to film. Course creators recording intros. Founders sending weekly updates. Anyone who makes short explainer or social clips and is tired of the lighting-and-retake routine for 20 seconds of footage.
 
What you need:
✓ A computer and a phone. The flow genuinely uses both.
✓ A paid Google AI plan on a personal account.
 
What's still rough:
⚠ Not available in the EEA, Switzerland, or the UK.
⚠ English only.
⚠ Generating video with your avatar sits behind a paid plan for now.
⚠ It's your likeness. No recording other people. No minors.
 
 
The walkthrough
 
Step 1: Open the avatar recorder
 
Selecting Avatar in the Gemini Add files menu
 
Start on your computer. Go to gemini.google.com and find the text box at the bottom, the one you normally type prompts into.
 
Click Add files. In that menu, click More uploads, then click Avatar. A QR code appears on screen.
 
 
 
Step 2: Record your face and voice
 
Pick up your phone or tablet and scan the QR code with the camera. It opens the recording flow and asks for camera and microphone access. Allow both. Then follow the on-screen instructions to record your face and voice.
 
This takes about 2 mins. Google checks the recording before it accepts it, and a few specific things get it bounced:
 
• Hold the phone at eye level, not down below your chin.
• Use light that isn't too dim or too bright. Face a window if you can.
• Keep your eyes, nose, and mouth visible. Glasses are fine.
• Record somewhere quiet, with no other voices anywhere in the background.
• Check what's behind you. No other people, and no photos or posters with faces in them.
 
Watch out: The background-faces rule is the one that catches people. A family photo on the shelf, a poster, someone walking through the room, any of those can fail the recording. Record against a blank wall and you remove the most common reason for a retake.
 
Take this seriously. The recording is the source of every video you generate later, so a careful 3 minutes here saves you a second recording session afterward.
 
 
 
Step 3: Finish and confirm on your computer
 
Go back to your computer. The recording syncs across from your phone on its own.
 
You get a chance to look at the result first. If it's good, click Use Avatar and it's saved to your account. If something feels off, the lighting, the framing, the audio, click Retake and run the recording again. Retake stays available permanently, so you're never locked into your first attempt.
 
Don't rush this check. The avatar you confirm here is the one that represents you in everything you make next. It's worth 30 seconds of looking before you commit.
 
 
 
Step 4: Put your avatar into a prompt and generate a video
 
Adding your avatar to a Gemini prompt
 
Now the fun begins.
 
Back in the Gemini prompt box on gemini.google.com, you have 2 ways to call your avatar. Click Add files, then Avatar. Or type @ and pick your Google username from the menu that appears. Either way, your avatar drops into the prompt as @[your username].
 
Then write the prompt around it. Here's one to copy and adjust:
 
Create a 20-second video of @[your username] speaking to camera in a plain, well-lit room, saying: "Quick update on the launch. We're on track for the 15th, and I'll send the full brief tomorrow."
 
Watch out: This is generated video, so treat the first few as experiments. Likeness and lip-sync quality vary, and it's English only right now. AI still has quirks like my avatar sitting at a computer under the kitchen exhaust extractor - random!
 
A video clip generated with the personal avatar
 
 
 
Step 5: Find, retake, or delete your avatar
 
Your avatar lives in your Google account, and you should know how to manage it.
 
To find it: open the Gemini app on your phone, go to Settings, then Personal avatar. Retake is right there if you want to replace the recording.
 
To delete it: on gemini.google.com, go to Settings, then Personal avatar, then Delete. That sends you to the Selfie video page of your Google Account, where you delete the actual selfie and voice recording.
 
Here's the honest part, and it matters. Deleting removes your recordings from Google's systems and stops you from making any new content with that avatar. It does not delete videos you already made and published. Those stay exactly where they are.
 
So delete is not an undo button. It closes the door going forward. Anything you have already generated and shared is out in the world for good. Plan for that before you publish, not after.
 
Managing your avatar in Gemini settings
 
 
How this all connects
 
Strip it back and this feature does one thing. It turns "I need to be on camera" from a production job into a typing job.
 
That sounds small until you count how often the production job is what stops you. The weekly team update you keep meaning to record. The course intro you've rewritten 4 times and never filmed. The 15-second social clip that somehow eats an afternoon. Record the avatar once and each of those becomes a prompt instead of a shoot.
 
The catch worth keeping in mind: a reusable, convincing version of your face now sits in your account. That is useful and a little uncomfortable at the same time, and I think both reactions are correct. So a few habits are worth adopting from day one. Label AI-generated video as AI-generated when you share it. Watch every clip before it goes out. And remember step 5, because once a clip is published, deleting the avatar will not pull it back.
 
Used with those habits, this is one of the more practical things Gemini has shipped in a while. Without them, you have just made it very easy to put convincing words in your own mouth. The habits are the price of the convenience.
 
 
Try this now
 
Record your avatar today, even if you have nothing to use it for yet. The recording is the slow part, and getting it accepted is the only real friction. Clear it once while you have 10 free minutes and a blank wall, and it's banked.
 
Then generate a throwaway test clip with this:
 
Create a 15-second video of @[your username] saying hello and introducing this week's project.
 
Watch it back. You will learn more about what this feature is good for from a single real clip than from any description, including this one. Have fun with your AI avatar!
 
 
 

👉 Know someone drowning in AI news? Forward this to them or send your unique referral link

Cheers, Tim

Keep Reading