A term of notes, in one file.
Photographs of a whiteboard, a reading pack and a scanned handout. One searchable file at the end of it, without paying for storage or uploading a single page.
For students and researchers who work in whatever tool is already open.
Why it usually goes wrong.
The enemy is the forty-page phone scan where one page is upside down and two are half a table. Rotating and cropping before merging costs a minute and saves an hour of scrolling later.
If the reading pack has a text layer, converting it to Markdown is the fastest way to get quotable text into notes. If it does not, no amount of converting will invent one.
In this order
Every step runs in the tab
Fix the upside-down pages
Rotate the whole document or just the pages that came out wrong. The turn is written into the file.
Open Rotate PDFTrim the desk out of the frame
Set margins as percentages and watch page one shrink in the preview before saving. The crop is a window, not a cut — the trimmed content is still in the file.
Open Crop PDFBind the photographs
One image per page, at the image’s own proportions, so a portrait page is not stretched.
Open JPG to PDFPull out the text
Reads the text layer the file already contains and rebuilds headings, bullets and lists. A scanned page with no text layer comes out empty and is reported as such.
Open PDF to Markdown
- Handwriting is not read. The Markdown converter reads characters the file already contains; a photograph of notes has none.
- Heading levels are inferred from font size relative to the page median. It is a good guess and it is still a guess.
- Tables and multi-column layouts arrive as running text. A two-column paper converts badly and there is no layout mode to fix it.
- The first 300 pages of a document are converted; the rest are reported as skipped.