Google Vids adds personal AI avatars and Gemini Omni video editing

Google has evolved Google Vids from an AI-assisted workplace tool for presentations into a comprehensive video creation platform. The company announced updates on Thursday that enable users to generate personalized digital avatars and utilize the Gemini Omni model for conversational video editing.

The new capabilities remove the requirement for cameras, microphones, or physical sets. By submitting a selfie and a voice recording, users can create a digital version of themselves that looks and sounds like them to deliver scripted messages. According to Androguider, these avatars feature realistic lip-syncing and voice modulation, making them suitable for onboarding videos, product demonstrations, and corporate training.

Related YouTube video

Conversational editing with Gemini Omni

The integration of Gemini Omni, a multimodal model first introduced at I/O 2026, transforms the editing process. Rather than using a traditional timeline, users can now edit videos through natural language prompts. Digitaltrends reports that users can simply describe changes to Omni to adjust lighting, swap backgrounds, or add post-production effects.

Gemini Omni is designed as an all-purpose content model capable of processing text, images, sketches, voice recordings, and existing footage. This allows for a step-by-step refinement process where each new instruction builds on the previous one. This approach helps preserve the look, subjects, and settings across repeated edits, representing a departure from previous workflows where users often had to scrap an entire project and start from scratch after making an error. These tools are also compatible with footage recorded on mobile phones.

For visual generation, Vids leverages the Veo 3.1 model, which Androguider describes as Google's most sophisticated video model. Veo 3.1 produces human-like avatars that react naturally to scripts and can generate complementary video clips based on reference images and text prompts. For example, an avatar explaining passwords can be paired with a Veo 3.1 clip of a password being typed into a text box. These generated clips are currently limited to 8 seconds.

Granular control and accessibility

Google is providing users with more precise control over AI narration. Users can insert bracketed commands, such as [excitedly], directly into a script to alter the avatar's pacing, emotion, or add start-up sound effects. When a user types an opening bracket, the system suggests available steering tags. To simplify this, a Apply audio tags option can automatically insert these cues throughout a script.

These voice-steering tools are available in both Rapid Release and Scheduled Release domains. Supported tiers include:

• Google Workspace Business and Enterprise editions
• Education Plus and nonprofit organizations
• Individual users and personal Google accounts
• Google AI Pro and Ultra subscribers

The avatar feature is accessible at vids.new. Users can choose from 12 preset AI avatars with predefined voices—which can be previewed by hovering over the options—or create their own. While the interface supports local languages, the avatar feature is currently available in English.

Safety and competitive positioning

To prevent the misuse of synthetic media, Google has implemented several guardrails. Every AI-generated clip includes a SynthID invisible watermark to verify its source. Furthermore, personal avatars are tied strictly to the account holder's likeness and Google account. Access is restricted to users aged 18 or older in supported regions.

Cryptopolitan suggests these restrictions are a reaction to the failures of other synthetic video tools, such as OpenAI's discontinued Sora, which allowed the creation of videos featuring public figures. By implementing these controls, Google Vids now competes directly with specialized AI video companies including Synthesia, HeyGen, D-ID, and Captions.

Quick Start: Creating a Google Vid

Step	Action	Tool/Feature
1	Navigate to vids.new	Browser
2	Select "AI Avatars" from toolbar	Interface
3	Upload selfie and voice recording	Personal Avatar Setup
4	Input text in "Add your script"	Script Editor
5	Refine lighting or backgrounds via chat	Gemini Omni

Google is currently soliciting user feedback to refine the avatar technology. Users can rate generations as Good or Bad via pop-up windows or submit detailed reports, including screenshots, through the Help > Help Vids improve menu.