Taskbase Homepage Museum

Every homepage taskbase.com ever had, in one place. Visuals only. Forms, analytics and backends are dead or stripped out.

TIRA-1478.

Run it

./serve.sh          # port 4780
./serve.sh 4999     # or pick one

Then open the gallery. It lists every homepage with a preview.

Layout

sites/<slug>/       one homepage, static, index.html at the root
sites/<slug>/meta.json    era, stack, source, status  (see CONTRACT.md)
sites/<slug>/preview.png  gallery thumbnail
fetch/              scripts that pulled the sites in
tools/gallery.py    builds index.html from the meta.json files
tools/serve.py      static server, sites live under /s/<slug>/
HISTORY.md          the timeline

serve.sh regenerates the gallery every time, so adding a site is just: drop it in sites/, write meta.json, restart.

Where the sites come from

• wayback — mirrored from the Internet Archive with fetch/wayback_mirror.py.
• git — extracted from history with fetch/git_snapshot.py. Three places: the old Java/AngularJS webapp (2015–2018), the standalone code.taskbase.com/taskbase/homepage.git repo (2017–2018), and the homepage/ folder in this monorepo (2018–2023). See HISTORY.md.
• live — the current site, mirrored with fetch/live_mirror.py.

Taken out before this went public

• Elementor Pro bundles. 19 files across the three WordPress mirrors: five widget chunks (loop, loop-carousel, nav-menu, nested-carousel) plus Pro's own frontend, webpack runtime, element handlers and two stylesheets. Pro is paid software, so we cannot hand it out. Deleted. Cost: on 2024-06-wordpress-hello and 2025-05-wordpress-lms the carousels no longer scroll and the nav dropdowns no longer open. Both say so in their meta.json, and both are partial now. Elementor's free chunks are still here — those are fine to ship.
• Internal bits and pieces. A staging box IP, a hosting project slug, the analytics and HubSpot account ids, one password in old 2016 source, and a handful of internal Google Docs links. Either generalized ("an AWS eu-central-1 staging host") or replaced with REDACTED. Personal email addresses are redacted too; the public info@ / support@ ones stay.

Known limits

Old snapshots lose things the archive never crawled: some fonts, some images, video. JS-heavy sites lose animations and anything lazy-loaded. Anything that needed a node build is kept as source plus a specimen page, not a live render — status in meta.json says which is which.