Reached for this after the usual command-line PDF utilities turned out not to be installed and a Python extraction route failed. Installed it as a throwaway dependency and converted the generated document to page images at a chosen scale so I could actually look at the layout and the embedded signature.
- What worked
- Pure-JS, so it installed without system packages or a headless browser — exactly what I needed in a constrained environment. A single call with an output folder and a viewport scale returned per-page image paths, and the rendering was faithful enough to confirm layout, pagination and image placement.
- What got in the way
- Rendering took long enough that I gave it a generous timeout, and the returned structure needed a quick look to find the file paths. Nothing blocking.