Get your handwritten notes off paper

Recognising handwriting is one of the few things where the free tool on your phone genuinely beats the open-source one on a server. Apple and Google have both spent years on it, it runs in a second, and it costs nothing.

• It is better at handwriting. Not slightly — noticeably. Classic OCR engines are built for printed text and treat cursive as noise.
• It is instant and free. No upload, no queue, no credits.
• Your notes never leave your desk. Nothing is sent to us, and on iPhone nothing is sent anywhere at all — Live Text runs on the device.
• It handles the awkward cases — a whiteboard photographed at an angle, a page with a coffee ring, a margin note in a different pen.

The fastest way: Live Text

1. Photograph the page. Fill the frame, keep the camera square to the paper.
2. Open the photo in Photos and tap the text-selection icon in the bottom-right corner (a square with lines in it).
3. Tap and drag to select, or tap Select All, then Copy.
4. Paste it wherever you like — Notes, an email to yourself, or straight into the campaign importer.

For a stack of pages: Notes' document scanner

1. Open Notes, start a new note, tap the camera icon and choose Scan Documents.
2. Hold the phone over each page — it detects the edges, straightens the page and shoots automatically. Keep going for the whole stack.
3. Tap Save. You get one multi-page PDF, already deskewed and contrast-corrected.
4. Tap into the scan and use Live Text on each page to copy the text out.

Google Lens — the best free handwriting reader there is

1. Open the Google app or Google Photos and tap the Lens icon.
2. Point it at the page, or open a photo you already took.
3. Tap Text, then Select all, then Copy text.
4. Copy to computer is worth knowing about: if you're signed into Chrome on your PC, it pastes straight there.

Windows 11

• Snipping Tool — take a snip of the image, then click Text Actions and Copy all text. Built in, no install.
• PowerToys Text Extractor — press Win+Shift+T, drag a box over anything on screen, and the text is on your clipboard. Works on a photo, a PDF, a video still, anything.
• OneNote — paste the picture into a page, right-click it and choose Copy Text from Picture.

macOS

• Open the image in Preview (or just press Space in Finder) and select the text directly — Live Text works on the Mac too.
• For a folder of images, the Shortcuts app has an *Extract Text from Image* action you can point at a whole batch.

Anything, including a stack of scans: Google Drive

This one is free, handles handwriting, and is the easiest route for a multi-page scanned PDF:

1. Upload the image or PDF to Google Drive.
2. Right-click it → Open with → Google Docs.
3. Drive runs OCR and gives you a document with the original image at the top and the recognised text underneath it.
4. Select the text, copy it, and close the document without saving if you don't want to keep it.

If you want a repeatable, offline, scriptable route — a hundred pages of *printed* material, say — Tesseract is the standard open-source engine. Be realistic about it: it is very good at clean printed text and poor at handwriting.

# Windows:  winget install UB-Mannheim.TesseractOCR
# macOS:    brew install tesseract tesseract-lang
# Linux:    sudo apt install tesseract-ocr tesseract-ocr-ell poppler-utils

# One image -> text on screen
tesseract page.jpg - -l eng

# One image -> page.txt
tesseract page.jpg page -l eng

# Greek, or Greek and English mixed on the same page
tesseract page.jpg page -l ell+eng

# A whole folder
for f in *.jpg; do tesseract "$f" "${f%.jpg}" -l eng+ell; done

A scanned PDF

A scanned PDF is pictures of pages, so it has to be turned into images first. pdftoppm comes with poppler:

# Split the PDF into 300 DPI page images
pdftoppm -r 300 -png notes.pdf page

# OCR each one
for f in page-*.png; do tesseract "$f" "${f%.png}" -l eng+ell; done

# Stitch the results together in page order
cat page-*.txt > notes.txt

Most bad OCR is a bad photograph rather than a bad reader. In order of how much difference they make:

1. Flat, even light. The single biggest factor. A shadow across the page — including your own head — costs more accuracy than anything else on this list. Daylight near a window, or a lamp to the side rather than behind you.
2. Square to the page. Photograph from directly above, not at an angle. The scanner apps straighten it for you; a plain camera does not.
3. Fill the frame. The page should be nearly the whole picture. Cropping in afterwards throws away the detail you needed.
4. Flatten the paper. A notebook's curve near the spine is where recognition falls apart. Press it flat, or photograph one page at a time.
5. One page per shot. A spread photographed as one image gives you two half-resolution pages.

A few pages

1. Go to Import my campaign.
2. Paste the text straight into the box. That is the whole step.
3. Everything arrives as a draft you check before anything is saved, and one click undoes the entire import.

A notebook's worth

1. Save each session as its own .txt or .md file.
2. Name them with the date: s07-2026-05-22-the-old-shrine.md. The importer reads the number and the date out of the filename, so your sessions arrive dated and in the order you played them.
3. Add them all at once with the same button — or zip the folder and upload that.