At last year’s I/O keynote, Google announced the rollout of Flow, an AI creative studio that lets creatives use natural language prompts to generate AI videos. Utilizing a “for creatives, by creatives” mission, Flow is designed to streamline the content creation process — allowing you to edit, create, iterate and animate all in once space, eliminating the need to purchase multiple platforms — backed by Google’s most advanced AI models.
This year, Google announced several major updates coming to Google Flow and Flow Music — Google’s generative AI platform dedicated for producing music and creating songs — at its annual developer conference, Google I/O.
Elias Roman, senior director of product management at Google Labs, discussed the intersection of AI and creativity, and the company’s mission to implement AI tools for powerful storytelling and creative expression.
Read also: Google’s AI Filmmaker Program Flow Helped Creators Make 100 Million Videos
New features and updates coming to Flow
Roman outlined three creative demands that guided Flow’s latest updates: a desire for cross-modality (breaking media-specific silos), a need to defragment highly fractured creative workflows and a push for more precise control and character consistency. Roman highlighted that creators frequently navigate an expensive, disjointed ecosystem of single-purpose apps, which breaks their creative “flow state.” To address these common pain points, Flow is rolling out four main updates.
First, Flow is transitioning from just a prompt-and-output tool into a conversational agent powered by Gemini models that acts as an end-to-end creative co-pilot with full memory of past and current projects. For instance, Flow’s agentic AI can become more of a “sounding board” and brainstorming partner when it comes to deciding on the dialogue of a scene or where the plot should go next in your story.
Google is also launching “Flow Tools,” a feature allowing creators to use natural language to instantly code and share custom tools or workflows. With this feature, you can create specific tools, like a video resizer or shaders, without needing to know how to code. You can also choose to share any tool you made with other Flow users.
Additionally, Flow will integrate a new generative model called Gemini Omni Flash, unlocking precise, video-to-video conversational editing and robust character consistency, including the user’s own avatar. Google describes Gemini Omni Flash as the Nano Banana, but for making videos. This feature is now available in Flow to all global Google AI subscribers.
Finally, Google is unveiling native mobile apps for Flow and Flow Music (more on this below), which will launch to enable on-the-go brainstorming and creation. Flow is now available on Android in Beta (iOS coming soon) and Flow Music is now available on iOS (Android coming soon).
I was able to catch a live demo of Flow’s conversational agent by planning a decade-accurate, 1980s-themed Times Square story, complete with persistent specific constraints (like adding an “Easter egg” miniature pinscher dog into every image). He also demonstrated precise, side-by-side video editing using natural language commands, custom custom tools such as custom ASCII character video rendering and a personalized avatar system.
Flow is transitioning from just a prompt-and-output tool into a conversational agent powered by Gemini models that acts as an end-to-end creative co-pilot.
What’s upcoming for Flow Music
Google Flow Music, the platform for creating and sharing music, is expanding its catalog of features.
For starters, more granular, precise editing is coming to Flow Music. Now, you’ll be able to edit any single part of a song, such as editing lyrics, translating lyrics or reworking the beat without editing the rest of the track.
You’ll also now be able to create covers of your favorites tracks. With this feature, you can take your favorite full song — keeping the melody and structure — and change the style, like turning a pop playlist into a lo-fi study version of it.
Flow Music is also implementing its latest Gemini Omni Flash model to let you create and share music videos on the platform. You can use Omni Flash to conversationally prompt the model to guide the style, scenes and essentially “direct” a music video that matches the vibe of your song. This feature is now available to all Google AI subscribers.
Read also: Over 40 Million AI Videos Have Been Made With Google Veo 3 Since May: How My Expert Testing Went
