Skip to content

How This Site Is Built

Curious what happens between your Google Doc and the page on this site? Here is the whole pipeline.

The pipeline

flowchart LR
    GD[Your Google Doc] --> REG[Registry sheet<br/>doc name + Google link]
    REG --> SYNC[Python sync script]
    SYNC --> MD[Markdown pages]
    SYNC --> IMG[Full-res images]
    SYNC --> PDF[PDF file]
    MD --> SITE[This site]
    IMG --> SITE
    PDF --> R2[Cloudflare R2 storage]
    R2 --> SITE
  1. You write in Google Docs. The doc is the single source of truth — the site never holds content of its own.
  2. A registry sheet lists what's published. Each row is one document: its name, category, and a link to the doc. If it's not in the sheet, it's not on the site.
  3. A sync script converts each doc. It downloads the doc straight from Google in three formats:
    • Markdown becomes the web page you're reading. The script strips things that don't belong on a website — the letterhead, the doc's title line, and the doc's table of contents (the site generates its own live one from your headings).
    • DOCX is used only to recover your images at full resolution, since Google's web conversion shrinks them.
    • PDF is saved so every page can offer "View PDF" and "Download PDF" next to a link to the original doc.
  4. The site is rebuilt and published to nanodocs.pages.dev (Cloudflare Pages). Every page ends up searchable, readable on phones, and in step with the docs.

Where the PDFs live

The pages and images are hosted on Cloudflare Pages. The PDFs are not: Cloudflare will not accept a single file larger than 25 MB, and some tool SOPs exceed that. So the PDFs are stored in a Cloudflare R2 bucket (nanodocs-pdfs) and the site fetches them when you click View or Download. The buttons still look like ordinary links on this site; a small Pages Function maps /assets/pdfs/… to the matching file in the bucket.

On a laptop, the sync still writes the PDF next to the pages so local preview works without the cloud. After a doc's PDF changes, that file also has to be uploaded to R2 or the live View PDF button stays on the old copy. Details and commands are in the repo README.

What this means for you

  • Edit the doc, never the site. Site files are overwritten on every sync.
  • You only edit the doc. The sync extracts the page, images, and PDF. Getting a new PDF onto the live site is a short upload step for whoever runs the repo (see the README), not something authors do in Google Docs.
  • Readers can always reach the original: every page links to a read-only preview of your Google Doc.

What this means for your readers

  • Readable anywhere. Pages adapt to any screen — a phone at the tool, a desktop in the office — far more comfortably than a PDF.
  • Searchable. Every page is indexed, so one search box covers all SOPs and policies.
  • Always current, still printable. Pages stay in step with the source docs, and a PDF copy is one click away when paper is better.