I would like to request screen sharing capabilities for Gemini Live. This would allow it to spot issues in other tabs that screenshots or video recordings might miss. Yes, I know frame extraction is possible, but having Gemini Live share the screen would be something never seen before; no other system—not even Anthropic, OpenAI, or DeepSeek—has this feature. Showing the AI the shared screen in real-time would be a unique selling point.It would significantly improve performance and speed for projects and programming tasks.
he system already works very well with mobile devices and AI Studio; the only modification needed would be to disable real-time voice output, allowing Antigravity to simply “see” and respond via the interface—whether in the Antigravity 2.0, IDE or the CLI. Just remove the voice and stick to text-based interaction.Removing the voice and relying on text, you reduce the GPU load and focus entirely on real-time visual analysis.While still giving the user the option to enable voice if they wish.