This does exactly what it says: pulls text from a PDF, lets you provide translations in a JSON file, then rebuilds the PDF with your translated text while keeping the original colors, backgrounds, and layout intact. The two-script workflow is straightforward. You extract strings, translate them however you want, then apply them back. It supports the usual suspects for fonts: Latin scripts, simplified and traditional Chinese, Japanese, and Korean. The output preserves styling better than most translation tools that just give you plain text or reformatted documents. Good for when you need a translated PDF that actually looks like the original instead of a mess.
npx skills add https://github.com/wshuyi/translate-pdf-skill --skill translate-pdf