Audio capture.
Every browser already ships a microphone and the code to record it. Yet when you want to capture a quick voice memo, you end up on some free site that watermarks the audio, caps you at sixty seconds, or quietly uploads your voice to a server you've never heard of. For a feature the browser gives away for free, that's a strange tax. This blueprint builds you a recorder that runs on your own page, your voice stays on your machine, and the download button actually works.
A voice memo app that captures audio straight from the microphone using the browser's own recording APIs, draws a live waveform while you talk, and gives you playback plus a download of the finished clip. No account, no upload step, no length limit someone else decided for you. The recording happens in the tab and the file lands in your downloads folder.
You watch it happen in the task workspace: the agent scaffolds the project, wires up the microphone capture and the waveform canvas, starts the dev server, and confirms the page records and plays back before handing you a live preview. The code lands on a branch in your repository with a pull request ready.
The first version is a base you steer with plain follow-up messages:
Each follow-up wakes the same repository and the agent continues on its own code.
Do I need to know how to code? No. You describe the recorder you want; the agent writes it, verifies it captures and plays audio, and ships the code to your repository.
Does my audio get uploaded anywhere? No. Recording and playback happen in your browser tab, and the download is a local file. Nothing leaves your machine.
Will it work without an internet connection? Once the page is loaded, the recording runs locally, so a dropped connection won't interrupt your memo.