I Work in Healthcare Tech. Here's Why I Built a RAG Tool for Clinical Documents.

I didn't set out to build a RAG application. I set out to solve an annoying problem I kept watching happen. The Problem With Most AI Demonstrations in Healthcare Temperature 0. No creativity. Deterministic responses only. The model either finds the answer in the document or it says it doesn't know. Explicit system prompt. The model is told directly: answer only from the provided context. If the answer isn't there, say so clearly. Do not guess. How RAG Actually Works (Without the Hype) ↓ (at query time) Question Building It: The Decisions That Mattered Chunking strategy What I Learned Building This The Code What's Next A few things I want to add: • Local LLM support via Ollama — so it works without an API key and keeps documents entirely on-premises. For clinical use cases, data leaving the building is a concern. • Multi-document querying — ask a question across a folder of documents at once • Better pre-processing — handling scanned documents with variable quality, which is very much a solved problem in OCR but needs connecting to the retrieval pipeline If you work in healthcare tech and have thoughts on what would make this more useful, I'd genuinely like to hear from you.