You can now clone yourself into a video without ever turning on a camera. Google Vids just added a tool that builds a talking, gesturing digital version of you from one selfie and one voice clip, and it ships alongside a second update that lets you edit any clip just by typing what you want changed.
Here’s the thing. Google didn’t announce this quietly. Two days after TechCrunch reported on OpenAI’s Sora shutting down back in March, Google went ahead and shipped the exact kind of feature that made Sora both wildly popular and deeply unsettling to a lot of people: an AI double of yourself that can say things you never actually recorded. Only this time it comes wrapped in a Google account, an age gate, and a watermark.
Let’s break down what actually changed, how to use both new tools, and where this puts Google against the avatar apps you’ve probably already heard of.
What Google Actually Added to Vids
Google Vids started life in 2024 as a fairly modest tool for building workplace videos inside Google Workspace, the kind of thing you’d use for a training clip or a project update. On July 16, 2026, Google rolled out two updates that push it well past that original job description.
1. Gemini Omni video generation and editing. Omni is Google’s multimodal model that turns text, images, and audio into video. Inside Vids, you type a plain-language prompt, optionally attach a reference photo or a rough sketch, and Omni blends those inputs into a generated clip. It also handles step-by-step edits on existing footage, so you can ask it to swap a background, fix bad lighting, or add an effect to a video you shot on your phone, without starting the whole project over.1, 2
2. Personal avatars. Upload one selfie and one short voice recording, and Vids builds a digital version of you that can deliver a script you type out, no camera and no re-recording required. Google says every avatar is tied to the account holder’s own likeness and locked to their Google account, so you can’t build one of a coworker or a stranger using their photo.1, 2
I’ll be honest, the timing is the interesting part here. Google explicitly frames this as a controlled alternative to what Sora tried to do in the open, and the guardrails it built in tell you exactly which failure mode it’s trying to avoid.
How to Create Your AI Avatar in Google Vids
If you want to actually try this instead of just reading about it, here’s the process end to end.
- Confirm you’re eligible. Personal avatars are limited to users 18 and older, and only in certain regions at launch. Open Google Vids and check whether the avatar option appears in your account before planning anything around it.2
- Check your subscription. Both new features require a Google AI Pro or Ultra plan, or a Google Workspace business account. A free Google account will not show these tools.2
- Upload your selfie and voice sample. Inside Vids, look for the new avatar creation option and follow the prompts to upload one clear photo of your face and record a short voice sample. Good lighting and a quiet room make a real difference in how natural the result sounds.
- Write your script. Type out exactly what you want your avatar to say. Short, conversational sentences tend to read more naturally than long, formal ones.
- Generate and review. Vids renders your avatar delivering the script. Watch the full clip before sharing it. Small mismatches between speech and expression are the easiest tell that a video is AI-made, and you’ll want to catch those yourself first.
- Export and share responsibly. Every clip carries an invisible SynthID watermark, so anyone can verify it was AI-generated. That’s a built-in transparency check, not an optional extra.2
How to Use Gemini Omni to Edit a Video With Just Words
The editing side of this update matters just as much as the avatars, maybe more if you already shoot your own footage.
- Start with a prompt or a clip. Type what you want to see, or upload footage you already shot on your phone.
- Add a reference image if you have one. A photo or even a rough sketch gives Omni more to work with than text alone.
- Let Omni generate the first draft. It mixes your prompt and any images into a first version of the clip.
- Describe your edit in plain language. Want a different background? Say so. Want the lighting fixed? Just describe the problem. No timeline scrubbing, no manual color correction.
- Keep editing step by step. Because Omni supports iterative edits, you can keep refining the same clip instead of regenerating it from scratch every time you change your mind.1, 2
What this really means for anyone who has fought with a timeline-based editor at 11pm trying to fix one bad clip: the friction that used to require actual editing skill just turned into a conversation.
Who Can Actually Use This Right Now
Access is narrower than the headlines suggest, so check this before you get excited about a specific use case.
- Gemini Omni in Vids: Available to Google AI Pro and Ultra subscribers and Google Workspace business customers.2
- Personal avatars: Same subscription requirement, plus an 18-and-older age minimum and a region restriction that Google hasn’t published as a full list.2
- Free Google accounts: Not included in this rollout.2
If you’re already paying for AI Pro and mostly using it for research and writing, this is effectively a free unlock. If you’re not subscribed, video generation and a personal avatar are a fairly compelling reason to reconsider, especially if you were already eyeing Veo access.
Google Vids vs the AI Avatar Apps You Already Know
Vids isn’t the first place you could build a talking AI avatar. HeyGen, Synthesia, Captions, and D-ID have been doing versions of this for a while, mostly aimed at marketers and course creators. Here’s how the pitch compares.
| Tool | Best for | Where avatars live |
|---|---|---|
| Google Vids | Workspace users who want quick, personal video updates without extra software | Bundled into Google AI Pro/Ultra and Workspace business plans |
| HeyGen | Marketing teams building interactive, conversational avatar videos | Standalone subscription product |
| Synthesia | Corporate training and course content at scale | Standalone subscription product |
| Captions | Creators making short-form social video with an AI presenter | Standalone app, mobile-first |
| D-ID | Developers building avatar features into their own products | API and standalone product |
Google’s actual advantage isn’t better avatar quality. It’s distribution. If you already live inside Gmail, Docs, and Drive, an avatar tool that’s already sitting in your Workspace account beats signing up for one more subscription just to send a quick video update.
The SynthID Question, and Why Sundar Pichai Won’t Be Getting the Sam Altman Treatment
When Sora launched, one of the first things people did was make Sam Altman say things he never said. Google clearly built this feature with that exact scenario in mind. Personal avatars are locked to the account holder’s own likeness, meaning you cannot upload a photo of Sundar Pichai, or anyone else, and put words in their mouth.1
Every clip, avatar or otherwise, carries an invisible SynthID watermark that lets anyone verify the video was AI-generated.2 That’s a meaningfully different starting point from a tool that let strangers generate videos of a sitting CEO, and it’s probably the single biggest reason Google feels comfortable shipping this now rather than a year ago.
The Honest Limitations
- Google hasn’t published the full list of countries eligible for personal avatars at launch, so availability may be inconsistent depending on where you are.2
- An invisible watermark verifies a video after the fact. It does nothing to stop someone from believing a fake in the moment, before anyone thinks to check.
- Avatar quality from a single selfie and a short voice clip will vary. Don’t expect broadcast-quality lip sync on your first try.
- This is a fast-moving rollout. Region access, pricing, and feature limits are the kind of details Google tends to adjust within weeks of a launch like this.
The Bottom Line
Google just turned Vids from a workplace video tool into something closer to a full AI video studio, and it did it by watching exactly what went wrong with Sora and building the opposite defaults: locked to your own likeness, tied to your account, watermarked by default. If you’re already paying for Google AI Pro, open Vids this week and try building one video with your own avatar before you decide whether this is a gimmick or a genuine time-saver.
If you know someone who spends their week recording the same “quick update” video over and over, send them this. Their next one might not need a camera at all.
Sources and further reading
Just Being Resourceful! 
Leave a comment