Web2026Live product
Movie & TV Discovery Platform
A live discovery platform that turns 1.3 million titles into a personal place to discover, track, and find where to watch.
- Platforms
- Web
- Focus
- Ask naturally · 1.3M+ titles · Discovery to decision
Project view
Visit the live project (opens in a new tab)The story
The problem
Streaming offers endless choice, but finding something worth watching often means bouncing between services, lists, and review sites.
What it delivers
Cinelas brings discovery, rich title information, personal lists, viewing progress, and detailed ratings into one responsive product anyone can use.
At a glance
Ask naturally
Search by mood, theme, plot, actor, or half-remembered scene, not just predefined filters.
1.3M+ titles
Movies and shows stay searchable in one catalogue.
Discovery to decision
Move from a suggestion to streaming providers, custom lists, viewing progress, and detailed ratings.
Built with
- RAG
- .NET
- React
- Azure
- Docker
- pgvector
- TypeScript
- PostgreSQL
- Tailwind CSS
- Azure OpenAI
Grounded recommendations (RAG)
Cinelas uses retrieval-augmented generation (RAG): it searches 15,000 prepared movie and show records by meaning, then gives the closest matches to Azure OpenAI to choose and explain. The AI works from real catalogue results instead of guessing.
Reliable, controlled AI
Cinelas removes titles already watched and checks every suggestion against its catalogue, preventing invented results. Sign-in, usage limits, caching, and monitoring keep the live service reliable and its costs predictable.
A closer look
Continue exploring
More projects
- Web2026Online BookingA complete booking journey for customers, with a live operations view for the people running the business.
- Web2026Personal PortfolioThe site you are using now: a privacy-friendly, accessible portfolio self-hosted on a Raspberry Pi at my home.
- Systems & Integration2025Newspaper Advertising PlatformA simpler way for newspaper subscribers and local businesses to place adverts, while staff keep customer information accurate behind the scenes.