Three apps built in a weekend with Gemini 3: a virtual try-on tool using hand tracking, a multiplayer art game where Gemini scores paintings for a global gallery, and a music platform that turns songs into draggable vinyl records. Vote on which one deserves a full tutorial.
I built 3 apps in 1 weekend using Gemini 3, Google's newest and most powerful model that's not just good at coding but also design and animation. So let's see how it performed. The first idea I had was a virtual try-on app where you can upload a screenshot of any clothing item and try on instantly in real time. Gemini was able to accomplish this by integrating the MediaPipe library, which is an open-source ML framework that enables tasks like facial detection, hand gesture tracking, and texture mapping. You can even use your finger to pinch an item and move it around. The second app is a multiplayer art economy game where the artist would draw a painting based on a prompt, which is then evaluated by Gemini 2 0.5 flash for similarity. If the score is larger than 60%, the painting can be listed through the gallery where users globally can vote to change the price and also purchase them. This was my first time making a game besides my CS assignments, which was a bit less fun. The last one is the interactive music sharing app where you can turn songs you want to share with your friend into record discs, which can then be played once you drag it onto a vintage record player, and this page can be turned to a link to share it. Overall, I think Gemini did a really good job in the design and also integrating a lot of Google's own ML models. But it also strikes me that AI doesn't just make coding faster. It also allows more personalized, creative, and fun software. If you want my prompts, comment "prompt," I'll send it to you. Also, please let me know which app you want me to deploy to production so you guys can use it and also make a full tutorial on.
