MedVextract is an AI-powered system, to aid VetRec extracting veterinary SOAP notes, follow-up tasks, medication instructions, client reminders, and veterinarian to-dos from consult transcripts. Built with BAML for LLM-powered extraction, FastAPI for the backend, and React for the frontend, it supports Practice Management System (PiMS) integration (e.g., Ezyvet), HIPAA/SOC 2 compliance, and scalability for multi-clinic veterinary practices.
- SOAP Note Generation: Automatically generates structured SOAP notes (Subjective, Objective, Assessment, Plan) and client-friendly discharge summaries.
- Task Extraction: Extracts follow-up tasks, medication instructions, client reminders, and vet to-dos from transcripts.
- Customizable Templates: Supports VetRec-style template selection for SOAP notes.
- PiMS Integration: Exports notes to systems like Ezyvet.
- Multilingual Support: Processes transcripts in multiple languages (e.g., English, Spanish).
- Scalability: Handles high traffic with Celery, Redis, PostgreSQL, and Kubernetes.
- HIPAA Compliance: Encrypts sensitive data and avoids logging PHI.
- Interactive UI: React frontend with VetRec-inspired blue/white design.
- Swagger Documentation: API documentation at
/docs.
- Backend: FastAPI, Python, Celery, Redis, PostgreSQL
- Frontend: React, TypeScript, Material-UI
- AI: BAML with GPT4o-mini for LLM-powered extraction
- Infrastructure: Docker, Kubernetes (optional)
- Python 3.11+
- Node.js 18+
- Redis
- PostgreSQL
- Docker (optional for containerization)
- Clone the repository:
git clone https://github.com/dotnetspark/medvextract.git cd medextract/backend