See what 5 builders are making with Gemini Omni
Google’s Gemini Omni AI model enables video creation and editing through conversational commands, with developers showcasing its capabilities in projects ranging from multi-angle filming to animated transformations.
Google introduced Gemini Omni, an AI model designed to simplify video creation and editing through conversational commands. The tool allows users to adjust camera angles, replace objects, and convert sketches into animations while maintaining realistic physics and coherence. Developers gained access to Omni, which combines real-world knowledge with intuitive editing features to produce high-quality videos from text, images, or audio references. Projects demonstrated include multi-perspective city scenes and dynamic environmental changes like shifting from day to night lighting.
Builder Leon Lin captured a single subject from over 20 angles across varied urban settings, showcasing Omni’s ability to generate diverse perspectives without losing scene continuity. The model’s physics-based understanding ensures realistic transitions between shots, such as zooming in or out while maintaining the subject’s position. This capability reduces the need for multiple camera setups or reshoots, streamlining the production process for creators and filmmakers seeking complex visual sequences.
Carlos Santana’s project demonstrated Omni’s voice-controlled editing, transforming an outdoor scene by altering lighting, weather, and seasonal elements like snow cover. The model’s real-time adjustments allow users to experiment with visual styles without manual post-production work. By interpreting natural language commands, Omni enables rapid iteration, making it useful for content creators, advertisers, or educators who need to adapt visuals for different contexts or audiences.
Pan’s work highlighted Omni’s sketch-to-video feature, where hand-drawn elements like lemons or scissors were animated into whimsical scenarios such as submarines or sharks. Jerrod Lew’s demo further illustrated style transfer, rendering a single clip in live-action, anime, and claymation styles while preserving the subject’s movement. Hyperagent’s projects included visualizing park redesigns, personifying data with animated explanations, and gamifying task lists, demonstrating Omni’s versatility across creative and practical applications.