Google has revealed Flowa new artificial intelligence platform designed for those who work in the field of film production, both professional and amateur level. Announced during the conference Google I/O 2025Flow stands out for direct integration with three of the most advanced models developed by Google: Veo for the video generation, Images for the creation of images e Gemini for text processing. This new tool represents a step forward compared to previous experiments such as Videofxgoing to constitute a real ecosystem for visual narrative. Designed to facilitate creative expression, Flow allows you to produce cinematographic clips starting from simple descriptions in natural language, guaranteeing high loyalty to prompt and an interesting narrative consistency between the different scenes.
The system allows you to control in detail elements such as angles of the camera and movement movement, and also offers advanced editing tools such as Scenebuilderwhich allows you to expand a scene already created without losing continuity. With Flow TVGoogle also proposes a dynamic video gallery generated by the AI, offering concrete and technical examples from which to draw inspiration. Currently, Flow is Available in the United States through paid plans Google to the pros and ultrabut the company includes a global extension in the coming months. We know more closely Flow and then let’s compare it with other top tools of the competition, such as Sora by Openii, Dream Machine of Labs Lobs e Runway by Runway Ai.
The characteristics of Google Flow, the tool to create videos with the AI
Flow It is an advanced video creation platform based on artificial intelligence, designed specifically for creatives and filmmakers. Built around the most advanced models of Google – Veo (for the video generation), Images (for the generation of text images) e Gemini (for the understanding and generation of natural language) – Flow allows you to develop cinematographic scenes and clips with visual coherence and high quality. Below we briefly summarize some of his Main technical characteristics:
- Deep integration with veo: It allows you to obtain realistic and cinematographic visual results, with high loyalty to physics and settings.
- Simplified prompt thanks to Gemini: You can describe your vision in natural language, without the need to give complex commands.
- Modular management of “ingredients”: Each resource created (characters, environments, objects) can be reused in multiple scenes with continuity and consistency.
- Advanced direction checks: Includes professional tools for control of the camera (movements, angles, prospects).
- Scenebuilder: It allows you to extend or modify a scene while maintaining narrative and visual continuity.
- Asset management: allows you to organize and preserve resources and creative prompts for efficient use.
Flow is the evolution of the previous videofx experiment and is currently accessible through subscriptions Google Ai Pro (with 100 generations per month starting from 19.99 dollars/month) and Google Ai Ultra (with early access to Veo 3 and native audio generation starting from $ 249.99/month) in United Stateswith expansion expected in other countries in the coming months.
It should also be noted that Google is actively collaborating with directors to integrate flow into their creative processes, collecting feedback to further improve the platform.
The differences with Sora, Dream Machine and Runway
Compared to Flow, Sora by Openi adopts an approach more focused on the creation of short clips – generally between 5 and 20 seconds – starting from simple textual prompts. Although visually impressive, Sora has some limitations, in particular in the generation of realistic human faces, for reasons related to safety and moderation. Its distinctive features include Remix, which allows you to edit existing videos keeping the structure, and “Storyboard”, useful for building visual narrative through the insertion of text or images in specific points of the clip. At the level of accessibility, Sora is integrated into the chatgpt plans, with a basic version of $ 20 and a Pro version of $ 200 per month, which offers more duration and video resolution.
Dream Machinedeveloped by Lobs, is particularly appreciated for the speed of generation and the photojournalist quality of the results. Thanks to the Ray2 model, it improves the reproduction of movements and physical behavior of objects, making the scenes more credible. Another innovation is represented by the “boards”, useful for organizing creative projects, and by the possibility of maintaining coherent characters in more content starting from a single reference image. The platform also offers a mobile app, thus expanding the possibilities of use on the move.
In the end, Runwaywho arrived at his third generation with the Gen-3 Alpha modelallows considerable customization. Users can insert images in specific points of the timeline and use advanced commands such as outpainting, which adapts the video format (vertical/horizontal) without losing visual consistency. The platform has already been adopted in the professional sector, with real applications in films and music video clips. A prominent element is “frames”, the generator of images that allows you to precisely define the visual style and to replicate it on more content.
