I'm building camera marking for printed worksheets

I'm building a feature that lets you mark a printed worksheet just by photographing it. It isn't open to everyone yet, but here's what I'm making.

Why I'm building it

Marking is quietly heavy work.

Twenty problems a sheet, checked one at a time against the answer page. When you're tired you make mistakes, and by the time you hand the wrong answers back, your child has moved on to something else. I've stalled at exactly that point myself and let the day's practice quietly go unfinished.

If marking took a few seconds, you could do only the problems that were wrong, right there. No need to redo the whole sheet. That's what I'm trying to build.

What it will do

Your child writes their answers on a printed worksheet, and you photograph it. The QR code in the corner identifies which worksheet it is, so there's nothing to select — no grade, no topic. You get circles, crosses, and a score on screen. If a box was read incorrectly, you can fix it on the spot.

There's no app to install. It's built to run in your browser, on the diglabo site.

What happens to the photo

The handwriting is read on diglabo's servers, so the photo you take is sent to diglabo. It is not processed only on your device.

Before the feature opens to everyone, I'll set out how the images are handled, including how long they're kept, in the privacy policy.

Where it stands

The hardest part is how accurately handwritten digits can be read. A 0 that looks like a 6, a 1 that looks like a 7 — in a child's handwriting. That accuracy is what I'm working on now.

This started as a smartphone app, but I've consolidated it into something that runs entirely in the browser. Skipping the "install an app first" step makes it far more likely to actually get used at home.

There's no release date yet. I'll announce it here once there is one.