Glimpse

TV & Movie discovery app with AI‑assisted recommendations, streaming availability, and fast, cached detail pages.

Live: https://glimpse.lucasanderson.dev

Features

• Search people, movies, and series with rich, media‑forward results
• Title pages with seasons/episodes, cast and crew, trailers, reviews, and recommendations
• Streaming availability via JustWatch, alongside TMDB metadata and ratings
• AI recommendations from short “mood/vibe” prompts
• Watchlists (schema + server wiring) and authenticated, protected routes
• Fast loads with Server Components, ISR, LQIP placeholders, and Redis caching

Stack

• Frontend: Next.js 14 (App Router, Server Components, ISR), React, TypeScript, Tailwind CSS, Radix UI/shadcn
• Backend & APIs: Node.js, Edge/Serverless routes, Zod validation, OpenAI API
• Data: PostgreSQL (Supabase), Drizzle ORM
• Caching: Upstash Redis (KV REST)
• Auth: Supabase Auth (JWT, RLS), SSR session handling
• Infra: Vercel (Edge Runtime), Docker (local), GitHub Actions (scheduled backfill)

Architecture Overview

• App Router with colocated server components in src/app
  • Media routes under src/app/(media) for movies, TV, people
  • Edge API for AI ranking at src/app/api/recommend/route.ts (model: gpt-4o-mini)
• Data layer with Drizzle ORM (src/db) targeting Supabase Postgres
  • Tables for movies, tv_shows, profiles, watchlist, watchlist_items, and summaries
• Caching & performance
  • Upstash Redis via src/services/cache.ts for LQIP blur data, popular scores, and AI result payloads
  • Next/Image with placeholder blur and image domain rewrites (next.config.mjs)
  • ISR for detail pages plus on‑demand revalidation
• Security
  • SSR Auth via Supabase and protected routes in src/middleware.ts
  • Bot filtering allow/block lists and simple IP/route rate limiting (150 req/min) in src/lib/rateLimit.ts

Getting Started

Prerequisites

• Node 20+ and npm

Install

npm install

Environment

Create .env.local and provide the following (use your own keys/secrets):

# Required
NEXT_PUBLIC_SUPABASE_URL=...
NEXT_PUBLIC_SUPABASE_ANON_KEY=...
POSTGRES_URL=...                             # Supabase connection string
TMDB_ACCESS_TOKEN=...                        # TMDB v4 access token
KV_REST_API_URL=...                          # Upstash Redis REST URL
KV_REST_API_TOKEN=...                        # Upstash Redis REST token

# AI recommendations (optional locally)
OPENAI_API_KEY=...

# Revalidation/backfill
REVALIDATE_SECRET=...
SITE_URL=http://localhost:3000

Run dev server

npm run dev

Scripts

• npm run dev — start Next.js in development
• npm run build — build for production
• npm run start — start production server
• npm run lint — run ESLint
• npm run backfill — populate Redis LQIP, DB summaries, popular people scores, and warm caches
• npm run revalidate — on‑demand revalidation of key pages (home, discover, popular IDs)

GitHub Actions runs a scheduled “Backfill and Revalidate” workflow (.github/workflows/backfill.yml).

Caching Details

• Redis namespaces
  • lqip:{movie|tv}:{id} — BlurData for images
  • popular:people — Precomputed popularity score payload
  • match:{sha1(tags)} — AI recommendation results (24h TTL)
• Edge API (/api/recommend) caches AI output for 24h and filters for complete metadata before returning results.

Security & Rate Limiting

• Supabase SSR auth with protected routes: signed‑in users redirected away from landing; unauthenticated users redirected to /signin for protected pages
• Bot handling: aggressive blocklist and allowlist for image optimization bypass
• Simple per‑IP, per‑route rate limiting: 150 requests/minute

Deployment

• Vercel with output: "standalone" and Edge runtime for the AI API route
• Image config allows image.tmdb.org and images.justwatch.com; /tmdb/:path* rewrites to TMDB CDN
• Set all environment variables in Vercel Project Settings and connect Supabase and Upstash

Attribution

This product uses the following data sources:

• TMDB — Movie & TV metadata
• JustWatch — Streaming availability

These services are not endorsed or certified by either provider.