Video by _radnolan

I am a Figma fanboy, but paper just came out, so let's give it a try. All right, we got some familiar hotkeys. Design or build anything with your favorite AI agent. Try this prompt. Full of clout. We are booping. It's dilly-dallying. Oh, there you go, okay. Yeah, it's making something now. So look at that, it's actually creating code and then putting it into the design. There we go. Got a status bar. Man, that animation of it coming in is super dope. Let's really quickly just pop this open, see what we can change. So we got all the layout built out. That's cool. Can we make components? Let's pull this out. Component. Is there nothing for components? That's kind of fucking crazy. Instead of running Claude to design something specifically, I'm running Claude to build something. I'll pull those into a Figma design file. What's really impressed me about paper is that the output from Claude was considerably better. It required me a lot more work on Claude's end to get it to something good that I wanted to pull into Figma. Because right now, Claude is really good if you give it a lot of context and plans and thinking. But if I just said, give me an expense tracker, the fact that paper came up with something way better than Claude out of the box shows that's a big feature for paper is that outputs are good. Where it's making it from Claude, sending it into Figma, getting it created and all that. I feel like Figma's gotta be coming for that, what paper's got, where you don't even need to do that step of sending it back and forth. It just is in the design side. But until that happens, it still works. You can still do it.