Google has rolled out a major update to the Gemini app, introducing a smarter and easier way to create AI-generated videos. With this new feature, users can upload up to three reference images - called visual “ingredients” - to guide the app while generating videos through Veo. These images can include characters, objects, styles, or entire scenes, allowing users to build more accurate and personalised videos.

This update focuses on solving three big problems:
• Character Consistency: Characters now appear the same across all scenes.
• Style Transfer: The video can adopt lighting, textures, and artistic styles from your reference photos.
• World-Building: Every object and scene in the video can match your custom world.

By using visual ingredients, users no longer need long, complicated prompts. The update is rolling out today and will be fully available next week for Google AI Plus, Pro, and Ultra subscribers. Google has also refreshed Gemini’s Tools menu to clearly show the model used for video generation, currently Veo 3.1.

This upgrade marks another big step in making AI video creation faster, simpler, and more creative.