Been building this solo for the past few months. It’s a movie tracker (watchlist, ratings, favorites) with an AI chat that does semantic search — describe a plot you half-remember, it finds the movie, instead of the usual “type the exact title” search everywhere else uses.
Stack: NestJS + Postgres/pgvector for the RAG search, React frontend. Deliberately skipped LangChain to keep full control over the embedding/retrieval pipeline.
Still pre-revenue, just started getting the first users this week. Would love feedback on:
• Does the AI search actually find the right movie, or does it hallucinate?
• Is a freemium model (unlimited basic tracking, capped AI searches/day) the right call here, or would you rather see it differently?
Happy to answer anything about the build too.