MySchool
An Android and Windows app to scan your class notes and find them filed.
MySchool scans class notes with the printer, reads the identifier written at the top of the page and files each sheet by subject, theme, chapter and lesson. It also turns them into revision sheets.
The binder, taken literally
I write my notes by hand on loose-leaf paper kept in a binder. MySchool turns them into a digital archive without changing that habit: you scan the sheet, the app files it in the right place, and you find it on your phone as well as your computer.
A code on every sheet
Each sheet carries a short identifier at the top, outside the margin:
HIST - TH1 - CH1 - S1 - 2627
Subject, theme (optional), chapter, lesson (or ANN1 for an appendix) and school year. The page number stays in the margin. The app follows the same hierarchy as the binder’s dividers: year, subject, theme, chapter, lesson or appendix, then the sheet with its front and back.
From sheet to archive
- Scanning in colour at 300 dpi, started from the app.
- Straightening the page and cropping it to its edges, before any reading, otherwise the identifier becomes unreadable.
- Reading only two areas: the top strip and the left gutter. It is much faster than reading the whole page, and more reliable.
- Checking: the form opens pre-filled and you confirm or correct it. If nothing is found, it opens empty rather than wrong.
- Filing: the image goes to private storage, the details to the database.
A scanner with no driver
The printer has no development kit for Android. It does speak eSCL, the protocol behind AirScan and Mopria: a small HTTP API, the same from Windows and Android. MySchool finds it on the local network and drives it directly, with no driver and no server in between.
Revision sheets
You pick a lesson, a chapter or a theme, and MySchool gathers the text of every matching sheet to produce a revision sheet ready to read, written by Claude. The result is saved, so it can be read from the phone too.
Technical choices
A single Flutter codebase for Android and Windows. Reading uses what each system provides offline: ML Kit on Android, the OCR built into Windows. Metadata lives in Supabase Postgres and images in private Supabase storage, protected by login: the free tier is enough for tens of thousands of pages. The interface borrows from loose-leaf paper, with its ivory sheet, ruling and red margin.
Gallery
