I built *GeminiDesk*, a free, open-source desktop client for Gemini (built with Qt/C++ under the hood) that runs on Windows, macOS, and Linux.

The goal was to move beyond the basic web wrapper and build a native experience focused on power-user features for productivity:

* *Keyboard Shortcuts:* Full control over your workflow using only the keyboard (switching models, opening/hiding the chat window). * *Advanced PDF Export with LaTeX:* The app uses its own renderer to export chats directly to a PDF file that correctly handles and renders complex mathematical formulas written in LaTeX. This is a huge productivity boost for research and technical conversations. * *Scheduled Deep Research:* You can schedule long, intensive deep research queries to run automatically at a later time. * *Audio Notification:* Get a non-intrusive sound alert when a lengthy response is finished generating, so you don't have to keep staring at the screen.

It's completely free and open source on GitHub. I'd love to hear your feedback on the features and performance.

What other quality-of-life features do you think are missing from the current Gemini web experience?

Thanks for checking it out!