Neverclick is a desktop application that I created when I was suffering from RSI. It uses local computer vision to detect clickable UI elements on your screen so that you can use your keyboard to perform mouse actions.
Most software in this space relies on accessibility APIs but Neverclick uses computer vision. This means that it works in every app. Additionally, it also produces really good bounding boxes around text so it's great for editing text and coding. And counterintuitively, it runs significantly faster than the traditional accessibility api approach.
It's 100% offline. It's not a wrapper around ChatGPT or Claude or any other online AI service. The computer vision system runs on your computer. I've optimized it is as best as I can, the hints should show up instantaneously even on 10 year old hardware. You don't need a powerful gpu to run it.
Also I swear on my life that it wasn't "vibe coded", I've been working on this since long before the recent AI revolution.