FPL Scraper is a Next.js dashboard and data pipeline for Fantasy Premier League transfer decisions powered by YouTube sentiment signals.

• FPL transfer recommendations (BUY / SELL / HOLD)
• Dashboard with KPI cards, recommendation explorer, event timeline, and video diagnostics
• Settings and sync controls via API routes
• Supabase-backed data access
• shadcn-style UI primitives with Tailwind CSS

• Next.js 16
• React 19
• Supabase
• Tailwind CSS + shadcn-style components

npm install

Create .env with the required keys (Supabase, YouTube, and any app secrets used by your setup).

npm run dev

Open http://localhost:3000.

• npm run dev - start development server
• npm run build - production build
• npm run start - run production server
• npm run test - run all tests
• npm run test:ui - run UI tests
• npm run smoke:supabase - smoke test Supabase routes

• /api/settings
• /api/sync/run
• /api/strategy/run
• /api/strategy/template
• /api/strategy/team
• /api/recommendations
• /api/runs
• /api/events
• /api/videos

• src/app - Next.js app routes and layout
• src/components - dashboard and reusable UI components
• apps/web/src - app service layer and APIs
• packages/pipeline/src - pipeline clients and scoring logic
• test - unit, contract, integration, UI, and e2e tests

• Keep .env local and out of source control.
• Deploy preview builds to Vercel for validation before production.

The app combines three data lanes into one decision dashboard:

1. Pipeline sync ingests YouTube videos, attempts transcript extraction, summarizes sentiment, and writes recommendation rows.
2. Strategy refresh builds elite-manager cohort snapshots, template ownership, momentum deltas, and team-fit insights for your entry.
3. Dashboard API serves the latest run-backed recommendations plus strategy panels and health diagnostics.

Use this sequence when operating or troubleshooting production data:

1. Save settings (entry_id, channel ids) in the dashboard Controls panel.
2. Run Strategy Refresh first (/api/strategy/run) to populate template/team-fit tables.
3. Run Sync Now (/api/sync/run) to generate a fresh recommendation batch tied to a new run_id.
4. Validate in dashboard:
   • Strategy cards are populated (Template Pulse, Team vs Elite, Transfer Radar, Momentum).
   • Recommendation counts reflect the latest run.
   • Pipeline Health shows the run completed (not stuck in running).
5. Use Video Diagnostics for transcript failure reasons (missing_transcript, ytdlp_not_installed, etc.).

When deploying code that changes DB contract:

1. Apply /scripts/supabase-schema.sql in Supabase SQL Editor.
2. Verify setup:
   • npm run supabase:verify
3. Smoke critical routes:
   • npm run smoke:supabase
4. Deploy after verify + smoke pass.

This prevents runtime drift between app expectations and Supabase tables.