Google Flow Adds Agentic Editing and Character Consistency via Gemini Omni

Google DeepMindGoogle DeepMind

· Updated

Google updated its Flow creative studios with Gemini Omni Flash to enable precise video editing and stable character identities across scenes. By introducing an autonomous agent for batch editing and natural language tool creation, Google is shifting AI video from single-clip generation to a managed production workflow.

Google integrated its new Gemini Omni Flash model into Google Flow, its AI-native creative studio. The update introduces a dedicated Google Flow Agent that can brainstorm dialogue and perform batch editing—applying a single creative change across an entire collection of video assets simultaneously to maintain visual harmony.
Core model
Gemini Omni Flash
Music model
Lyria 3 Pro
New capabilities
Batch editing, character consistency, plot brainstorming
Custom tool creation
Natural language (no-code)
Mobile availability
Android beta (Flow), iOS (Flow Music)
Access
Google AI subscription required for Omni features

This shift addresses the consistency gap that has historically limited AI video to short, disconnected clips. By leveraging Gemini Omni's world understanding, the system can now preserve character identity and voice across multiple scenes. This moves the platform beyond simple generation toward a production environment where AI functions as a persistent partner.

You can now use natural language to vibe code bespoke Flow Tools, such as custom video resizers or glitch effects, which can be shared or remixed. While the Google Flow Agent is available to all users, the Gemini Omni Flash features and tool creation require a Google AI subscription. Mobile apps are now available in beta.

View the full update on blog.google
Google DeepMind
Google DeepMind
@GoogleDeepMind
X

Google Flow 🤝 Gemini Omni Create more cinematic stories with our latest model, which brings batch editing, improved character consistency and more. Here’s what else is new for @FlowbyGoogle → https://t.co/corY3RwY7t #GoogleIO https://t.co/usg0Sudiv9

40retweets325likes
View on X

Still wondering? A few quick answers below.

Gemini Omni Flash is a multimodal AI model integrated into Google Flow to improve world understanding and precise video editing. It allows creators to blend real-world inspiration with generated content conversationally. The model specifically addresses character consistency, ensuring that a character's identity and voice remain stable across different scenes in a project.

The Google Flow Agent acts as an autonomous creative partner that can reason through complex tasks. It assists with early brainstorming, dialogue suggestions, and plot recommendations. For production efficiency, the agent can perform batch editing to apply changes across multiple assets at once and organize files into collections with intuitive naming conventions.

Google Flow Tools allow users to create bespoke creative workflows using natural language instead of traditional coding. Creators can build custom image editors, video resizers, or specialized shaders by describing what they need. These custom tools can be shared with the community, allowing other users to remix them into their own creative projects.

Google Flow Music now offers granular, section-by-section editing control. Users can highlight specific parts of a song to rewrite lyrics or restyle a beat drop without changing the rest of the track. A new Covers feature also allows creators to transform the entire style of a song while preserving its original melody and structure.

The Google Flow Agent is available to all users globally, while advanced features like Gemini Omni Flash and custom tool creation require a Google AI subscription. For mobile access, a beta app is available for Android users over 18, and the Google Flow Music app has launched on iOS with an Android version coming soon.

Every HeadsUpAI update is written based on its original source and reviewed before it's published. Read our editorial standards →

Google

See all AI news & updates from Google →

Share this update