Movie-recommendation AI chat application using own vector DB. The React client sends a user's request to a Python FastAPI backend, which searches a PostgreSQL/pgvector movie catalogue and uses OpenAI to return grounded human-like recommendations.

See it live here: www.booksynth.com/movies

Official page: f4cio.com/moviesrecommendation

Structure

• F4CIO.MovieRecommendation.UiWeb - React 19 + Vite chat UI.
• F4CIO.MovieRecommendation.Backend - FastAPI monolith with API, business logic, and data-access layers.
• dataset.json, ImportDataSet.py, VectorizePlots.py - movie data import and embedding utilities.

Run locally

1. Copy F4CIO.MovieRecommendation.Backend/.env.example to .env and set OPENAI_API_KEY and DATABASE_URL.
2. Ensure PostgreSQL has pgvector enabled and the database is ready (after creating empty Movies db run /Database/FinalDatabase.sql to populate it).
3. Run start-backend.bat and start-web.bat.

The API runs at http://localhost:5100 (/docs for Swagger); the UI runs at http://localhost:5101.