FamBam Taiwan
LiveTrip planner for 14 people across 3 generations
Started: July 5, 2026 · ~1 evening for the site + ongoing survey coordination
AI:
Claude Code
Claude CodeStack:Next.jsReact
TypeScript
Supabase
Vercel
TypeScriptSupabaseVercelA private, password-gated family website for a 14-person Thanksgiving trip to Taiwan — 7 kids, 6 adults, and one grandma, flying out of SFO. The family reviews the roster, dates, and running decisions log (Taipei plus a second base, a possible Japan extension, a chartered minibus with driver), and each adult fills out one survey covering their crew — responses land in Supabase and feed the itinerary. The same operator playbook as the Yosemite planner, scaled up to three generations with a real feedback loop.
// highlights
- Password-gated Next.js site on Vercel — one shared family password keeps the roster and dates off the open web without auth theater
- Survey pipeline: one submission per adult covering their kids, stored in Supabase behind an isolated Postgres role, tallied on an admin results page
- Coordinates 14 people across 3 generations — group airfare quotes, a chartered ~20-seat minibus, and a base-city decision tracked in a decisions log
- Docs and site live in one repo — commit and push after every planning session, so the plan and the code version together
// takeaways
- A shared password is the right security for family logistics — full auth would be theater; the gate plus Vercel hits the convenience/privacy sweet spot.
- AI is genuinely useful for real family logistics, not just demos — the survey schema, itinerary drafts, and decision docs all compounded across sessions.
- Supabase Free with one small table and one isolated role covers 90% of what a from-scratch backend would — right-size the infra to the job.