Loomersidian
Loom + Obsidian. Turn a Loom recording into a Markdown timeline of screenshots and transcript, ready to hand to any LLM agent.
If you’ve ever tried to explain a UI to an AI coding agent in text, you know how slow it is. Screenshots one at a time, descriptions of layouts, circling things in follow-up messages. It works, but it’s the dumbest part of working with AI right now. A screenshot with an arrow saying “move this down here” beats a paragraph of description by a thousand-fold.
Loomersidian (Loom + Obsidian) takes a Loom video plus its transcript and pairs every spoken passage with the screen on display when you said it. The output is a single timeline.md and a folder of frames — drops straight into Obsidian for cleanup, then into Cursor, Replit, or whichever agent you’re handing it to.
You talk through what you want. The tool does the rest. I’ve gone from long back-and-forth sessions to dropping in one file and walking away — the agent gets 80–90% there on the first pass.