MedScribe: AI‑driven OCR and LLM tool to extract and structure medical records.
MedScribe helps clinics turn scanned PDFs or images of patient records into structured data. It first runs Tesseract OCR, then feeds the raw text to an LLM (Anthropic or OpenAI) that maps it onto a predefined clinical schema, and finally presents the results in a Streamlit dashboard for human review. The CLI and export options let users batch‑process files and store results as JSON, CSV, or SQLite. It stands out by merging confidence scoring and a simple UI, offering a ready‑to‑use pipeline for medical offices without building their own stack.
View on GitHub →jacfruit404/MedScribe