What changed
Mito, a company focused on democratizing AI-powered video creation, announced the use of Google Cloud's infrastructure to manage the production of a 120-second national TV spot for Movistar Plus+ during the World Cup.
The project required generating over 5000 assets, including 1000 videos, across 65 iteration rounds.
The solution involved a serverless, asynchronous architecture built on Google Cloud, with services like Cloud Run, Cloud Pub/Sub, and Cloud Storage playing key roles in managing media assets and enabling real-time collaboration.
What this means in practice
The production team used a structured canvas frontend that enforces a 'Rule of 6' to limit prompt variables across different creative elements. This approach prioritizes still-frame generation before video rendering to maximize cost efficiency.
Media assets are stored in Google Cloud Storage and delivered via Cloud CDN, with expiring key signatures ensuring pre-release client assets remain private.
The AI Director copilot translates natural language instructions into precise model parameters, maintaining context, lighting, and visual style across scenes.
Distributed teams and brand clients can work on the same canvas in real time, with new generations saved as separate assets to allow for parallel creative exploration.
Limitations and availability
The solution relies on Google Cloud's infrastructure, including Cloud Run, Cloud Pub/Sub, and Vertex AI, to handle the heavy computational demands of AI video rendering.
The platform is designed for enterprise-grade AI storytelling, with a focus on handling large-scale projects and ensuring real-time collaboration. The AI Director copilot uses models like Omni, Veo, and Nano Banana, accessed through Vertex AI, alongside custom system prompts.
The architecture is primed for future improvements, with the potential to reduce rendering times from minutes to seconds as new world models and personalized content delivery capabilities emerge.
The supplied announcement does not specify the exact release date or availability of the full platform.
