Show HN: SC Screen Recorder: a browser-based screen recorder with added layers

[rikroots]

I’ve been working on a small proof-of-concept screen-recording tool that runs entirely in the browser. It lets you capture screens or windows, add a talking-head camera feed, layer multiple targets, drag and scale them, draw annotations directly on the canvas, and record everything to a local MP4/WebM file. There’s no backend, no build step, and no external dependencies apart from MediaPipe and Scrawl-canvas.

The project is intended to be hackable rather than polished, and includes a keyboard-accessible UI and a step-by-step walkthrough. Code and demo are here:

GitHub: https://github.com/KaliedaRik/sc-screen-recorder

Feedback very welcome.

[jcalloway_dev]

Really cool project — the "no backend, no build step" constraint is genuinely hard to pull off well and you did it.

Couple things I'm curious about:

How does performance hold up when you're layering multiple capture sources simultaneously? That's usually where browser-based tools start to sweat.

Also — the "hackable rather than polished" framing is smart positioning. Are you thinking about specific use cases you want people to fork this toward? Because the talking-head + annotation combo immediately makes me think async demo videos for indie devs who can't afford Loom Pro, or lightweight sales prospecting tools. Would be interesting to see what the community actually builds with it.

One suggestion: a short screen recording (meta, I know) showing the layering and annotation in action would probably cut your "time to wow" significantly. The README explains it well but seeing the canvas manipulation live would land faster.

Nice work on this.