website

FlickBolt

On-demand video capture of physical locations. A two-sided marketplace where Customers request video of a place and Capturers (mobile creators) fulfill those requests, plus a Twitch-style Live mode where capturers broadcast publicly and viewers tip / subscribe.

Stack

Layer	Tech
Frontend	Jekyll → GitHub Pages (site/)
API	Cloudflare Workers (Hono router) (workers/api/)
Database	Cloudflare D1 (SQLite)
Storage	Cloudflare R2
Video	Cloudflare Stream (live + VOD)
Sessions	Cloudflare KV
Real-time	Durable Objects (workers/matching-do/, workers/live-channel-do/)
Background	Cloudflare Queues
Payments	Stripe Connect (Express)

GitHub Pages serves zero dynamic content. Every dynamic interaction is fetch() from the static page to a Worker endpoint that returns JSON.

Repository layout

flickbolt/
├── site/                  # Jekyll site → GitHub Pages
├── workers/
│   ├── api/               # main API worker (Hono)
│   ├── matching-do/       # Durable Object for matching (Phase 6)
│   ├── live-channel-do/   # Durable Object for live chat (Phase 9)
│   ├── shared/            # cross-worker helpers
│   └── migrations/        # D1 SQL migrations
└── .github/workflows/     # CI: jekyll.yml + workers-deploy.yml

Local development

Site (Jekyll)

cd site
bundle install
bundle exec jekyll serve --host 0.0.0.0 --port 5000 --livereload

API (Workers)

cd workers/api
npm install
echo 'JWT_SECRET="dev-secret-change-me"' > .dev.vars
npx wrangler d1 migrations apply flickbolt_db --local
npx wrangler dev

Phase status (this commit)

• ✅ Phase 0 — Empty deployable skeleton.
• ✅ Phase 1 — Jekyll routes, layouts, page-scoped JS, mobile nav.
• ✅ Phase 2 — Workers API with Hono, D1 schema, health/version, CORS, error handler. Auth scaffolding too (signup/login/refresh/logout/me) so Sprint 2 can begin without rework.
• ⏳ Phase 3+ — see attached_assets/Pasted--FlickBolt-Build-Specification-...txt.

Things you (the human) still have to do for Sprint 1

1. Create a GitHub repo named flickbolt and push.
2. Enable GitHub Pages → source: GitHub Actions.
3. Buy a domain (flickbolt.com) — optional but recommended for cookie-based auth.
4. In Cloudflare:
   • Run wrangler login
   • wrangler d1 create flickbolt_db → paste database_id into workers/api/wrangler.toml
   • wrangler kv:namespace create flickbolt-sessions → paste id into wrangler.toml
   • wrangler r2 bucket create flickbolt-media
   • wrangler secret put JWT_SECRET (use a long random string)
5. Add GitHub Actions secrets: CLOUDFLARE_API_TOKEN, CLOUDFLARE_ACCOUNT_ID.
6. Push to main. The two workflows deploy independently.

Acceptance for Sprint 1 (per spec):

• Visiting flickbolt.com shows the FlickBolt landing page. ✅ (locally on Replit)
• curl https://flickbolt-api.guillaumelauzier.workers.dev/health returns {"ok":true}. ✅ (live)
  • Custom api.flickbolt.com is optional and waiting on DNS.