Hi Product Hunt! I'm Jay, maker of Shadow and now Cakie. I just became a dad to twin girls. Best thing that's ever happened to me. But 16-hour days are over, so working efficiently matters more than ever. I use AI agents for most of my product work, and voice typing already made explaining ideas much faster. Something was still missing, though. A lot of what we communicate at work isn't text. It's the screen in front of us: the layout, the interaction, what happened after a click, the exact part we're looking at. With a teammate, I'd share my screen, point, and talk. The screen is simply part of the conversation. With AI, I was still squeezing all of that into a text box. "Make this Get started button bigger… no, not the one in the top bar, the second one, in the Team card." So I built Cakie. 🎙 Record your screen like a Loom for your agent. Talk as if a teammate is sitting next to you. 👆 Point or draw when you want to be exact. Circle a button, or mark the screen. 🎞 Cakie picks the moments that matter. A minute of video is thousands of frames. Cakie uses what you said to choose the few you meant, so your agent gets a transcript plus the right screenshots, not the whole video. 📋 Review, copy, paste. Remove any screenshot you don't want, then paste one prompt into Claude Code, Codex, Cursor, or any agent. I now use it for almost every prompt I send while building Shadow, and Cakie itself. It's free while in beta on macOS. If you spend your day explaining things to AI too, I'd love for you to try it, and I'd especially love to hear where it gets things wrong.

Cakie

Turn screen recordings into prompts for AI agents

Turn screen recordings into prompts for AI agents