Google has rolled out a major set of creative controls for Flow, its AI-powered filmmaking and video-generation platform, through the new Gemini Omni 1.1 Flash model. The update adds start-and-end frame controls, 1080p and 4K exports, and a lower-cost 360p draft mode designed to help creators test ideas before committing to higher-resolution generations. Google says all three features are available in Flow starting August 27.
The update is aimed at making AI-generated video more controllable and practical for production workflows. Instead of generating an entire clip repeatedly at full quality, creators can now establish the beginning and ending points of a shot, experiment with inexpensive 360p drafts and then upscale selected results to 720p, 1080p or 4K. Gemini Omni 1.1 Flash is also being made available to developers through Google AI Studio and the Gemini Enterprise Agent Platform, extending the same video-generation capabilities beyond Flow.
Google Flow Gets New AI Video Controls
Google Flow was launched as an AI filmmaking tool designed around Google’s generative video technology and has gradually expanded into a broader creative studio.
The latest update focuses less on simply generating video and more on giving creators control over how a shot begins, ends, develops and gets finalized.
The three headline additions are:
- Start and end frame controls
- 1080p and 4K exports
- 360p draft generation
These features address three common problems with AI video generation: maintaining continuity, controlling transitions and avoiding the cost of repeatedly rendering high-resolution clips.
Google Flow Update At A Glance
| Feature | Details |
|---|---|
| Platform | Google Flow |
| New model | Gemini Omni 1.1 Flash |
| Start frame | Supported |
| End frame | Supported |
| Draft resolution | 360p |
| Draft speed | Up to 60% faster |
| Draft cost | One-third of standard 720p generation |
| Final exports | 1080p and 4K |
| Draft upgrade | 720p download |
| Scene extension | Up to 40 seconds cumulatively |
| Video references | Up to 3 seconds |
| Availability | Google Flow users |
| Developer access | Google AI Studio and Gemini Enterprise Agent Platform |
Google says the new capabilities are rolling out in Flow and are designed to support both quick experimentation and more polished production work.
Start And End Frames Give Creators More Control
One of the most important additions is the ability to specify both the first and last frame of a shot.
Instead of asking an AI model to independently interpret an entire sequence, creators can provide the desired starting and ending visual states.
Gemini Omni 1.1 then generates the video transition between those two points.
Starting Frame
↓
AI Understands Scene
↓
Camera / Character Movement
↓
Transition Generation
↓
Ending Frame
This can be especially useful for camera movements, cinematic transitions and looping shots.
Start-End Frame Applications
| Use Case | Potential Benefit |
|---|---|
| Camera orbit | More controlled movement |
| Zoom transition | Better framing continuity |
| Push-in shot | Defined beginning and ending |
| Pull-back shot | More predictable composition |
| Scene transition | Better visual continuity |
| Looping video | Easier return to original composition |
| Character movement | More controlled trajectory |
Google says the feature is designed to generate continuous video between two keyframes, making it suitable for complex camera orbits, zoom transitions and seamless loops.
Gemini Omni 1.1 Can Extend Scenes Up To 40 Seconds
The new model also improves scene extension.
Gemini Omni 1.1 can analyze up to 10 seconds of prior video context when extending a scene. Earlier models referenced only the final second, according to Google.
The expanded context is intended to improve visual consistency and narrative continuity.
Creators can extend videos in 10-second increments up to a cumulative length of 40 seconds.
Scene Extension Workflow
Original 10-Second Clip
↓
Gemini Analyzes Previous Context
↓
+10 Seconds
↓
+10 Seconds
↓
+10 Seconds
↓
Maximum 40 Seconds Cumulative
This gives filmmakers more room to build longer sequences without repeatedly starting from scratch.
360p Draft Mode Makes Experimentation Cheaper
Another major addition is Flow’s draft mode.
Creators can generate previews at 360p before committing to a full-resolution render.
Google says 360p generations can be produced up to 60% faster than standard 720p generation and cost one-third as much.
That makes the feature particularly useful for testing prompts, compositions and camera movements.
Draft Mode Economics
| Generation | Resolution | Relative Cost | Speed |
|---|---|---|---|
| Draft | 360p | 1/3 of 720p | Up to 60% faster |
| Standard | 720p | 1X | Baseline |
| Final | 1080p | Higher | Higher |
| Premium final | 4K | Higher | Higher |
The cost comparison is specifically between 360p and Omni 1.1’s standard 720p generation, rather than a claim that 360p costs one-third of every available output resolution.
Creators Can Test Multiple Ideas Before Final Rendering
AI video creation often requires several attempts.
A creator might want to test:
- Camera position
- Character movement
- Lighting
- Composition
- Scene pacing
- Prompt wording
- Object placement
Rendering each variation at full resolution can consume both time and credits.
Draft mode changes that workflow.
Idea
↓
360p Draft
↓
Review
↓
Modify Prompt
↓
Second Draft
↓
Compare Versions
↓
Select Best Take
↓
720p / 1080p / 4K Final
This approach effectively separates creative experimentation from final production.
Flow Can Upscale Selected Videos
Once a creator is satisfied with a draft, the selected version can be downloaded at 720p.
Google also says Gemini Omni 1.1 can generate polished 1080p and 4K outputs for professional production.
This creates a workflow in which low-resolution drafts are used for exploration and higher-resolution outputs are reserved for final delivery.
Resolution Workflow
| Stage | Resolution | Purpose |
|---|---|---|
| Concept testing | 360p | Fast experimentation |
| Draft selection | 360p | Compare variations |
| Preview / download | 720p | Review quality |
| Standard final | 1080p | Social / digital production |
| High-end final | 4K | Professional workflows |
The ability to separate draft and final rendering could help creators manage AI video-generation costs more efficiently.
4K Export Targets Professional Creators
The introduction of 4K output is particularly important for creators producing content intended for professional editing and distribution.
Google says 1080p and 4K outputs are designed for high-end digital, social and broadcast editing workflows.
That expands Flow’s positioning beyond experimentation and short social clips.
Potential 4K Applications
| Content Type | 4K Value |
|---|---|
| Film concepts | High-resolution previews |
| Advertising | Product and campaign footage |
| Social media | High-quality master files |
| Music videos | Cinematic sequences |
| Corporate video | Professional presentation |
| Broadcast workflows | Higher-resolution source material |
| VFX concepts | Better detail for editing |
For professional users, higher-resolution source material can also provide more flexibility during post-production.
Gemini Omni 1.1 Adds Video Reference Support
The model also supports video references as part of multimodal input.
Creators can provide up to three seconds of reference video when generating a new scene.
The goal is to preserve visual context and character consistency while incorporating movement from the reference material.
Video Reference Workflow
Reference Video
↓
Character / Movement Analysis
↓
Prompt + Reference
↓
Gemini Omni 1.1
↓
New Generated Scene
This could be useful for applications such as choreography, character animation, product demonstrations and visual effects.
Character Consistency Remains A Key Focus
Maintaining character identity across generated scenes has been one of the biggest challenges in AI filmmaking.
Google’s broader Flow platform already emphasizes character consistency, and Gemini Omni is designed to preserve identity and voice across scenes.
The start/end frame feature adds another layer of continuity by allowing creators to define exactly how a shot should begin and conclude.
AI Video Continuity
| Challenge | Flow / Omni Approach |
|---|---|
| Character identity | Consistent character generation |
| Scene continuity | Extended video context |
| Beginning of shot | Start-frame control |
| End of shot | End-frame control |
| Camera movement | Prompt-based direction |
| Reference movement | Video references |
| Iteration | 360p draft mode |
The combination is intended to make AI-generated footage behave more like a controllable filmmaking tool.
Google Flow Is Becoming An AI Creative Studio
Flow has evolved significantly since its initial launch.
Google originally introduced it as an AI filmmaking tool built around Veo, with features such as camera controls, scene building and asset management.
The platform has since expanded with AI agents, custom tools and mobile capabilities.
Evolution Of Google Flow
| Stage | Major Capability |
|---|---|
| Initial Flow | AI filmmaking |
| SceneBuilder | Extend and edit shots |
| Camera controls | Direct camera motion |
| Gemini Omni | Conversational video editing |
| Flow Agent | AI-assisted planning and editing |
| Custom Tools | Natural-language workflow creation |
| Mobile app | Creation on the go |
| Omni 1.1 | Keyframes, drafts and 4K |
Google says Flow has expanded to more than 140 countries and is increasingly positioned as an AI creative studio rather than a single-purpose video generator.
Flow Agent Adds Another Layer Of Automation
The latest video controls arrive alongside Google’s broader Flow Agent experience.
The agent can help users brainstorm, create and edit projects.
It can also generate multiple variations, batch-edit assets and organize creative materials.
Flow Agent Workflow
Creative Brief
↓
Flow Agent
↓
Planning
↓
Generation
↓
Variations
↓
Editing
↓
Asset Organization
This means creators can increasingly use Flow as an AI collaborator rather than manually operating every part of the creative workflow.
Creators Can Build Custom Tools Without Coding
Google has also introduced Flow Tools, which allow users to create custom tools and workflows using natural language.
A creator could, for example, build a custom image editor, video resizer or visual-effects workflow without writing conventional code.
This makes the platform more flexible for creators with specialized production requirements.
Custom Tool Workflow
Describe Tool
↓
Gemini Builds Workflow
↓
Creator Tests Tool
↓
Modify With Natural Language
↓
Use In Production
↓
Share / Remix
The addition of customizable tools could eventually make Flow more like an AI-powered creative operating environment.
Gemini Omni 1.1 Is Also Available To Developers
The technology is not restricted to Flow.
Google is making Gemini Omni 1.1 Flash available to developers through Google AI Studio and the Gemini Enterprise Agent Platform.
Developers can use the model to build their own generative-video applications and workflows.
Gemini Omni 1.1 Distribution
| Platform | Availability |
|---|---|
| Google Flow | Yes |
| Google AI Studio | Yes |
| Gemini Enterprise Agent Platform | Yes |
| Gemini API | Yes |
| Developer applications | Supported |
| Enterprise workflows | Supported |
This could allow third-party creative applications to incorporate the same video-generation and editing capabilities.
Adobe, Figma And Runway Are Among Early Users
Google says several companies are already using Gemini Omni in production or integrating it into creative workflows.
Adobe has integrated Gemini Omni Flash into Adobe Firefly.
Google also highlighted use cases involving Figma Weave, GMI Cloud and Runway.
The companies are using Omni for applications ranging from video editing to creative experimentation.
Gemini Omni Ecosystem
| Company | Highlighted Use |
|---|---|
| Adobe | Adobe Firefly video editing |
| Figma | Creative workflows |
| GMI Cloud | Educational and explanatory video |
| Runway | Video generation and editing |
The presence of established creative-software companies suggests Google is positioning Omni as infrastructure for a broader generative-video ecosystem.
AI Video Competition Is Shifting Toward Control
The latest Flow update reflects a broader change in generative video.
Early AI video models competed primarily on visual quality.
The next phase is increasingly about control.
AI Video Competition
Generation Quality
↓
Character Consistency
↓
Camera Control
↓
Scene Extension
↓
Start / End Frames
↓
Reference Videos
↓
Drafting & Iteration
↓
4K Production
The more precisely users can direct a model, the closer AI video tools can come to fitting into conventional filmmaking workflows.
Google’s latest update is therefore less about simply producing prettier video and more about making generated footage predictable enough to use in a production pipeline.
Lower-Cost Drafting Could Improve AI Video Economics
Generative video remains expensive compared with ordinary text generation.
Each failed attempt can consume credits or computing resources.
By shifting experimentation to 360p, Flow can potentially reduce the cost of creative iteration.
For professional users generating dozens of variations, this could have a meaningful impact.
Traditional AI Video Workflow
Prompt
↓
Full Render
↓
Not Satisfied
↓
Full Render Again
↓
Not Satisfied
↓
Full Render Again
Flow Draft Workflow
Prompt
↓
360p Draft
↓
Review
↓
360p Draft
↓
Compare
↓
Final 4K Render
The second workflow can potentially reserve expensive compute for ideas that have already been validated creatively.
Mobile Creation Is Another Important Component
Google is also bringing Flow capabilities to mobile.
The Flow Android app is available in beta for users aged 18 and older, while Google says an iOS version is coming.
The company’s earlier Flow update positioned the mobile experience as a way for creators to work on projects while away from desktop systems.
Draft mode is particularly suited to mobile because creators can quickly test ideas without waiting for full-resolution rendering.
Google Is Targeting Professional And Casual Creators
The combination of low-cost drafts, AI agents and high-resolution output broadens Flow’s target audience.
Casual creators can experiment cheaply.
Professional filmmakers can use start/end frames and 4K exports.
Agencies can build customized workflows.
Developers can integrate Omni into their own applications.
Potential Flow User Base
| User | Potential Use |
|---|---|
| Casual creators | Quick AI videos |
| YouTubers | Social content |
| Filmmakers | Previsualization and scenes |
| Advertisers | Product campaigns |
| Agencies | Client content |
| Designers | Creative prototyping |
| Developers | Custom AI-video applications |
| Enterprises | Marketing and communications |
This broad positioning could help Google compete more directly with dedicated AI-video companies.
The Bigger Picture
Google’s latest Flow update shows that AI video generation is moving from simple prompt-to-video creation toward a more controllable production workflow. Gemini Omni 1.1 Flash adds start-and-end frame control, longer scene extension, video references, 360p drafting and 1080p/4K output, giving creators more control over both experimentation and final production.
The most important change may be the combination of these features rather than any single capability. Creators can cheaply test ideas at 360p, define important visual states with keyframes, use references for continuity, extend successful shots and then render selected work at higher resolutions. That workflow could reduce wasted generation credits while making AI-generated footage easier to incorporate into professional editing pipelines.
Looking Ahead
Google’s next challenge will be proving that these controls can consistently deliver predictable results across longer and more complicated productions. Higher resolution alone does not solve the core problems of AI filmmaking; creators also need reliable characters, camera movement, scene continuity and precise adherence to creative direction. Gemini Omni 1.1’s expanded context and start/end frame capabilities are designed to address several of those issues.
Competition among AI video platforms is likely to increasingly focus on workflow rather than image quality alone. With Flow combining Gemini Omni, AI agents, custom tools, mobile creation and high-resolution exports, Google is attempting to build an end-to-end creative environment around generative video. If the company can make the process fast, affordable and controllable enough for professional users, Flow could move from an AI experimentation platform toward a serious component of modern video-production workflows
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
