What I learned building a RAG notes app
Embeddings, vector search, reranking, and getting Flutter, Spring Boot and Python to talk to each other
Series
I got curious about how RAG really works, so I built Cortex, a notes app that answers questions from your own notes. This series is what I learned on the way: embeddings, vector search with pgvector, hybrid search and reranking, testing the answers, and keeping personal data safe. Flutter app, Spring Boot API, Python AI service.

Embeddings, vector search, reranking, and getting Flutter, Spring Boot and Python to talk to each other
What a 30-question test set caught in my notes app: a reranker reading the wrong part of a note, and a phone number reaching the model.
