Comic Archive Tools

comic-archive-tools

A toolkit for working with comic/manga page archives: packaging and extracting CBR/CBZ archives, converting between PDF/PSD/image formats, small image utilities, and a Claude Code skill that automates Photoshop lettering prep (text detection, cleaning, paragraph text boxes). Each piece below is independent — use whichever one you need.

Overview

Script / skill	What it does
Manga lettering pipeline (SKILL.md + scripts/)	Claude Code skill — turns manga/doujinshi page scans into layered, letter-ready PSDs: text detected and erased, a native Photoshop paragraph text box added per speech bubble.
tools/images_to_cbr.py	Packages a folder of images into a .cbr archive.
tools/cbr_to_images.py	Extracts .cbr/.cbz/.zip comic archives into image folders.
tools/pdf_to_images.py	Extracts PDF pages as JPG images.
tools/psd_to_jpg.py	Converts Photoshop .psd/.psb files to JPG.
tools/rotate_images.py	Rotates every image in a folder by a given angle.
tools/stretch_pngs.py	Stretches every PNG in a folder to exact dimensions.

---

Manga lettering pipeline (Claude Code skill)

Turns a folder of manga/doujinshi page scans into layered PSD files ready to hand to a letterer: the original text erased cleanly, and an editable Photoshop paragraph text box already sitting in place of each speech bubble/caption, pre-filled with placeholder text in a comic lettering font. No more hand-drawing a Type tool box over every bubble before you can start translating — open the PSD and start typing.

Output: what's in the PSD

Every page becomes a PSD with:

• Layer "Original" (bottom) — the untouched scan, for reference or to restore anything the cleaning step damaged.
• Layer "Cleaned" — the same page with all detected CJK text erased, and only text: stroke detections are gated by the model's own text-block boxes and text-line map, so stroke-like false pos