Inspiration Healthcare costs are confusing, especially when someone is already stressed about symptoms or a referral. We built CareCost to give people clarity earlier, so they can understand likely GI care paths and cost ranges before they commit.
What it does CareCost is a map-first GI care shopping experience. It helps users translate symptoms into the most relevant procedure path, then shows nearby providers and realistic out-of-pocket cost bands so they can compare options quickly.
How we built it We built the frontend with Vite and React, with a Leaflet-based map and Auth0 authentication. The backend is FastAPI with MongoDB for provider and pricing data, and we route AI calls through Lava to Gemini. The AI assistant asks focused follow-up questions, suggests the next step in a fixed GI tree, and maps to validated procedure buckets.
Challenges we ran into Our hardest challenge was balancing helpful AI guidance with strict reliability. We had to enforce strong guardrails so the model could not invent procedures, while still keeping the conversation natural, fast, and useful for real users.
Accomplishments that we're proud of We’re proud of delivering a full end-to-end flow from symptom input to map-based cost comparison. We also built a provenance-first system where the decision tree remains the source of truth, which makes outputs more trustworthy and easier to explain.
What we learned We learned that trust and clarity matter as much as model quality in healthcare-adjacent tools. Structured validation, transparent outputs, and simple next-step UX made a bigger difference than adding more complexity.
What's next for CareCost Next, we want to expand beyond GI into more specialties, improve insurance-specific estimate personalization, and add stronger explainability so users can see exactly why a path or estimate was recommended.
Log in or sign up for Devpost to join the conversation.