Better3
let's talk

website · research group · 2026

vee-lab.eu gives a research group a site it can update from a text file, with no framework to maintain

A static site for the Virus in Ecology and Evolution lab at Radboud University: research, team, publications, two databases and nineteen protocol pages, generated as plain HTML with a Python build step for SEO and structured data.

  • html
  • tailwind
  • python
  • json-ld
  • netlify
read the case study ↓

What it is

A research group needs one place that says what it works on, who is in it, what it has published and what it gives away. Most university pages cannot be edited by the group, and most site builders bury the content in a tool the next postdoc will not know. This site is the other way round: a folder of HTML pages, one stylesheet, and a short Python script that adds everything a search engine wants before deploy. Anyone in the lab can change a page with a text editor.

what is on it
fig. 01 — every protocol has its own URL, so a method can be cited by link

What is on it

Research. One page for the question, virus-host interactions in microbes, and four angles on it: evolution, ecology, biotechnology and computational biology.

Team. Current members, a “join the lab” card with the contact route, and former members kept on the page rather than deleted.

Publications. The bibliography from 2013 to 2026, each entry with journal, year and authors, linked to Google Scholar.

Resources. The Gmeta Toolkit and the ArchaeaHQ database, and the protocol library: nineteen walkthroughs for tools such as HMMER, MMseqs2, Foldseek, geNomad and ColabFold, grouped as foundations, methods and advanced.

tree site/
fig. 02 — pages are files; the build step rewrites only the marked blocks in each head

The build step

The pages are written by hand. What a search engine, a social card or a crawler needs is not: canonical URL, Open Graph and Twitter tags, JSON-LD for the organisation and its lead, the analytics snippet and the sitemap. A Python script generates these into marked blocks in every page head, so they are consistent across 26 pages and never edited by hand.

deploy
fig. 03 — the marked blocks are the only part of a page a script touches
page health
fig. 04 — no JavaScript framework; the only external scripts are the fonts and, after consent, analytics

Design decisions

  • Files, not a framework. The lab outlives any one maintainer; HTML in a folder is the format with the longest shelf life.
  • Generate the boring part. Metadata, structured data and the sitemap come from one script, so a new protocol page gets them for free.
  • Protocols as pages. Each method has its own URL and can be cited in a thesis or a methods section.

What is not done

Publications are maintained by hand; a fetch from ORCID or Crossref is the obvious next step. The images are PNG and could be served as WebP. There is no light theme.