Videos, PDFs and images
The installer prepares the tools Claude needs for everyday work, so you can ask for a video, a PDF or a batch of images in any session - no setup.
| Installed | Used for |
|---|---|
| Node.js 22 + Remotion | motion design videos written in React (skill video-motion-design) |
| Chromium (one for everything) | HTML -> PDF (maxxing pdf, skill pdf-document), video renders and Scrapling share the same browser |
| ImageMagick, ffmpeg | resize, convert, compress, thumbnails (skill image-edit) |
| poppler-utils | read and preview PDFs (pdftotext, pdftoppm) |
| Scrapling + Chromium | web scraping: static, JavaScript and protected pages -> Markdown, CSV, JSON (skill web-scraping) |
| zip, unzip, jq, ripgrep | archives (exports, imports), JSON, fast search |
Videos
Ask "make a 20-second vertical video announcing our summer sale". Claude creates a project
(maxxing video new <topic>/videos/<name>), writes the scenes, checks a few frames, renders an MP4
and gives you its path (open it in Files). Remotion is installed once in
/var/lib/claude-maxxing/tools/remotion; each project links to it, so it takes a second to create
and weighs a few MB.
PDFs
Ask "make a PDF quote for …" or "turn this report into a nice PDF": Claude writes an HTML page and
prints it with maxxing pdf page.html page.pdf (A4, backgrounds, your fonts).
Web scraping
Ask "collect the name and price of every product on this site into a CSV" or "tell me every morning
if this page changes". Claude uses Scrapling: scrapling extract get|fetch|stealthy-fetch URL out.md
for one page (plain HTTP, or a real browser for JavaScript pages), or a short Python script run with
scrapling-python for structured data over several pages. Results land in the topic (data/),
with the script, so the job can be re-run or scheduled. It respects robots.txt and goes slowly by
default; personal or login-protected data only with your explicit say-so. No MCP server is added
(it would cost context in every session): Claude calls the command line.
Installed in /var/lib/claude-maxxing/tools/scrapling (every package pinned with its checksum) and tools/browsers (Chromium, shared with videos and PDFs), ~750 MB.
Images and SEO
image-edit covers resizing, webp conversion, compression, crops, thumbnails and contact sheets,
one file or whole folders. seo-audit checks a website for real (robots.txt, sitemap, titles,
descriptions, headings, canonical, Open Graph, Core Web Vitals through PageSpeed) and writes a
prioritized report in the topic.
Check or repair
maxxing video status # Node.js, Remotion, headless browser, Scrapling
maxxing deps list # versions of every component
maxxing deps update # (re)install the tools, as root
maxxing doctor warns (never fails) when they are missing.