Audition Aggregator

Collects open audition listings from regional theater websites into one searchable place.

A scheduled scraper fetches each theater's audition page, hands the text to Claude for structured parsing, and writes the results to Supabase. The Next.js app reads from there.

Stack

Next.js 16 (App Router, Turbopack) · React 19 · TypeScript · Tailwind CSS 4 · Supabase · Claude API

Setup

npm install
cp .env.local.example .env.local   # then fill in the values
npm run dev

Run supabase/schema.sql in the Supabase SQL editor to create the tables, then npx tsx scripts/seed-theaters.ts to load the theater list.

Environment variables

Variable	Used by	Purpose
NEXT_PUBLIC_SUPABASE_URL	browser	Supabase project URL
NEXT_PUBLIC_SUPABASE_ANON_KEY	browser	Read-only access (public read is intentional)
SUPABASE_URL	server	Same URL, server side
SUPABASE_SERVICE_ROLE_KEY	server, scraper	Writes; bypasses RLS
ANTHROPIC_API_KEY	scraper	Parsing audition pages
GITHUB_TOKEN	server, scraper	Files issues for feedback and broken pages
ADMIN_KEY	server	Gates the admin pages

Commands

npm run dev      # dev server
npm run build    # production build + typecheck
npm run lint     # ESLint

npx tsx scraper/index.ts                  # full scrape
npx tsx scripts/test-scraper.ts [url]     # smoke test one page

Layout

app/              routes: /, /past, /theaters, /add-theater, /admin/*
components/       UI
lib/              Supabase client, types, expiry rules, admin session
scraper/          fetch → parse (Claude) → write; entry point is index.ts
scripts/          seeding and one-off migrations
supabase/         schema.sql and seed data

Admin access

The admin pages are gated by ADMIN_KEY. Sign in at /admin/login; this sets an httpOnly session cookie good for 30 days. Once signed in, edit links appear next to each theater on /theaters.

The cookie stores an HMAC derived from ADMIN_KEY, not the key itself, so it can't be read back out. There is no key in any URL — earlier versions used a ?key= query parameter, which leaked into browser history and server logs.

Deployment

The app deploys to Vercel from main. All seven environment variables above must be set for both Production and Preview.

The scraper does not run on Vercel. It runs as a GitHub Action (.github/workflows/scrape.yml) daily at noon UTC, and can be triggered manually from the Actions tab. It needs SUPABASE_URL, SUPABASE_SERVICE_ROLE_KEY, and ANTHROPIC_API_KEY as repository secrets.

Database permissions

service_role needs UPDATE on theaters — the scraper deactivates theaters whose pages 404, and the admin page edits them. Without the grant these writes fail with permission denied for table theaters. See the bottom of supabase/schema.sql.