Inspiration
Confusing medical bills and language barriers make it difficult for patients to recognize and challenge overcharges.
What it does
MedBill compares bills with insurance statements, flags potential overcharges, answers multilingual questions, and drafts dispute letters.
How we built it
We combined React, FastAPI, Python billing checks, document extraction, and a local Ollama language model.
Challenges we ran into
We balanced local model speed, multilingual accuracy, and the need to preserve billing amounts during translation.
Accomplishments that we're proud of
We built a working prototype that connects billing findings to document evidence and helps patients take action.
What we learned
Useful AI needs reliable data, clear explanations, and checks that prevent it from inventing financial details.
What's next for MedBill
Improve response speed, expand billing-rule coverage, strengthen document extraction, and validate translations with native speakers.
Built With
- css
- fastapi
- llm
- natural-language-processing
- numpy
- ollama
- pydantic
- pymupdf
- python
- qwen3
- react
- reportlab
- rest
- tailwind
- typescript
- vite