Textsnap
Pull clean text from images and screenshots with a pip install and no GPU
Emerged finds new products worth a look, daily — what’s new today
Tenderness
Render documents whose layout you already know, so labels come for free
If you train document-reading models, this gives you pixel-exact bounding boxes by construction instead of paying for noisy OCR-derived labels.
A Python library from Paperchase Labs that renders documents from text and images with Cairo and Pango, recording where every block landed. Output goes to image, SVG, PDF or NumPy arrays, laid out by a flexbox-style engine with rich typography and composable pipelines. Because the structure is known before rendering, the training labels are exact and the runs are deterministic. Apache-2.0, installs with pip, though it needs Cairo, Pango and PyGObject on the system.
More like this
Penmentor
Pen your maths in Xournal++, press a key, get teacher-style ink marks back
FrontBack Scanner
Free flatbed photo scanning that finds, crops, pairs and stitches your prints
Dieline Studio
Free print-ready dielines for bags, pouches and boxes, sized in the browser
Docx-to-Markdown
Word to Markdown, page by page, using a real Word-compatible layout engine
P
PrintSheet
Lay out photos, labels and PDF pages on A-size paper and get a print-ready PDF