The most underutilized AI API is Gemini 2.5 pro with video input. This app is for our first ever intern, Kevin, who makes content. So I built an app that analyzes his videos before he posts. And it doesn’t just analyze the script. It literally “watches” the video in... seconds and provides feedback on visuals + dialogue. Side note: The agent now searches the web, and implements any API just give it your key.show more

@vibecodeapp How on earth did I miss that Gemini 2.5 Pro allows video input :O Also in the Gemini app or only via API?

@vibecodeapp It works on Google AI Studio which I use instead of Gemini app. @OfficialLoganK is video input in the Gemini app?

@vibecodeapp video in is really cool explored using it for interviews, finding signs of lying via body language, but it’s not so great and hard to do few shot with video vision in models is generally slept on imo by both labs and public

@vibecodeapp my father wants the ai to help diagnose difference between his golf swing on the driving range vs the course. can i get beta access to vibe code this app for him? video analysis comparing swings in each location is the answer.

@vibecodeapp Great app idea

@vibecodeapp ok this is 🔥but where did you get the audio from 😭

@vibecodeapp I have other skills bro

@vibecodeapp Building some crazy stuff with this atm. 💯

@vibecodeapp agree with you. pretty sure this is because the docs are hard to work with too. need a way simpler way to build multimodal / live realtime apps with Flash etc

@vibecodeapp Classy