Pages with rich text, nested under other pages, are a genuinely well-scoped build. Fast search that stays relevant across years of institutional knowledge is where real depth lives.
Borderline
Effort: a weekendRunning cost: $0/mo to start
Why CanICodeThis says Borderline
Pages with rich text, nested in a tree structure, is a well-documented pattern shared with this register's Confluence-style build spec
Version history (view and restore an earlier revision) is a straightforward append-only log
For a small team's wiki, this is squarely a weekend project
The parts that'll cause problems
Search that stays fast and relevant across years of pages needs real indexing, not a database text match
Space-level permissions and granular page restrictions matter more at company scale than they first appear
Deep integration with other tools (embedding live data from elsewhere) is real, separate integration work
How I'd build this
frontend
Next.js
database
Supabase
auth
Supabase Auth
hosting
Vercel
MVP scope
Pages with rich text, nested in a tree
Full version history per page
Full-text search across all pages
Simple shared login for the team
Postpone to v2
Granular per-page permissions
Live embeds from other tools
Comment threads on pages
What it actually costs to build
AI agent — with a subscription (Claude Pro/Max, Cursor, etc.)
$0 marginal
AI agent — pay-per-use API, no subscription
$43–$86 one-time
Hosting, once it's running
$0/mo (free tier)
Domain name, if you want your own
~$12/yr
Existing tools solving a similar problem typically run $5.75–$12/mo — see how they compare below.
The build prompt
Build me a team wiki: pages with rich text, nested in a tree structure, with full version history. A sidebar tree navigator and full-text search across all pages using a real search index, not a database LIKE query. SQLite or Supabase behind a small API, simple shared login for a small team.