Website / image-processing workflow
PrintableLineArt
An archived colouring-page website with a Python workflow for generating printable PNGs and PDFs, and a Laravel application for importing, organising and publishing them.

About PrintableLineArt
PrintableLineArt was a website for downloading colouring pages. I built the Python side that prepared the artwork and the Laravel side that turned those files into a browsable catalogue.
Getting an image from a model was one step. It still needed clean black outlines, a sensible page size, a printable file, a caption and somewhere to go. Doing all of that by hand would have made each new batch a small publishing job of its own.
The Python workflow used Stable Diffusion through Diffusers, with style and LCM LoRAs. It converted the results to black and white, checked how much of each page was filled with black pixels, and could try another seed when an image came out too dark. The standard outputs were US Letter PNGs at 300 DPI and raster PDFs. Optional exports included WebP and AVIF for the browser, plus SVG and vector PDF through Potrace.
Each image got its own folder and a metadata file containing captions, tags and the details needed for publishing. The files could then be placed in Backblaze B2 for the Laravel importer to pick up.
That importer handled the less photogenic work: finding the right assets, checking source references and file hashes to avoid repeat imports, sorting out slug collisions, and creating the catalogue entries. A small admin area let me edit the text and tags, change publication status and schedule pages to go live.
For visitors, the site stayed simple: browse a category, open a page, download the PNG or Letter PDF. WebP and AVIF variants were used for browser previews where available.
The original site is retired. I am keeping the project here because the work behind it is still relevant: Python image processing, structured metadata, background imports and a publishing application that connects them.
























