Watchlist

A personal movie and TV show tracking app with smart recommendations. UI generated with Claude Code.

Features

• Track movies and TV shows you want to watch
• Mark items as watched and rate them
• Multi-user support (cookie-based, no auth)
• Smart recommendations based on your watch history
• "Watching with" mode for group recommendations
• Autocomplete search powered by TMDB

Tech Stack

• Backend: Go with Chi router
• Database: SQLite
• Frontend: HTML templates, Tailwind CSS, DaisyUI
• APIs: TMDB (search, recommendations), OMDB (ratings)

Running Locally

1. Clone the repo
2. Copy .env.example to .env and add your API keys:
   • TMDB API key
   • OMDB API key
3. Run the server:

   go run .

4. Open http://localhost:8080

Recommendation Algorithm

Signals

The engine combines three data sources from TMDB for each watched item:

1. Similar titles - Movies/shows with similar themes, genres, and style
2. TMDB recommendations - Collaborative filtering based on user behavior
3. Keyword matching - Boosts candidates that share keywords appearing in 2+ watched items

Scoring

Each candidate accumulates points:

• +1 for appearing in a similar titles list
• +1 for appearing in a recommendations list
• +1 for each shared keyword (if that keyword appears in 2+ watched items)

Items need a minimum score of 2 to be included. Final ranking multiplies the score by TMDB popularity.

Categories

Results are sorted into four buckets:

Category	Criteria
Popular	TMDB popularity >= 30
Newest	Released in last 2 years
Hidden Gems	Popularity < 30, but score >= 2
Classics	15+ years old with popularity > 10

Group Mode

When using "Watching with" to get recommendations for multiple users:

• Items are weighted by how many users watched them (shared favorites score higher)
• Anything already on anyone's watchlist or watched list is excluded

Performance

• Concurrent API calls with goroutines
• 1-hour TTL cache for TMDB responses
• Semaphore limits parallel keyword fetches to 20