Nobody should have to rewatch a 30-minute call to find out what they're supposed to do.

So I built Loom to Tasks. Paste in a meeting transcript and it pulls out the decisions, tasks (with owner and deadline) and blockers in seconds. Python + Gemini with structured outputs.

The rule I cared most about: it never guesses. If nobody said who owns a task, it writes "unknown" instead of inventing a name. Tested against hand-written answer keys and passed 67/67 checks.

Code's on GitHub: github.com/AbrarShuvo/loom-to-tasks

If your team lives in meetings, I can set this up for your workflow.

Writing "unknown" when nobody named an owner is the rule I like most: an invented name is worse than an empty field. My RAG demo follows the same idea and says so when the guide has no answer. In your 67 checks, how many test the "unknown" cases, where the transcript is not clear?

Behind the scenes of shipping a client web app: the launch-day cut list is where the real product gets decided. This week we shipped with three features deliberately left out, not because they were hard, but because nobody on the client's team could explain who would use them or when. Every feature you ship is a promise to maintain it, support it and explain it to the next new hire. The apps clients keep paying for are not the ones with the longest feature list, they are the ones where every single thing on screen earns its place. Ship less, on purpose, then add only what real users ask for twice.

Ship less on purpose, love that. Half the features I've built got clicked once, by me, while testing 😅