RAG
Sehat Guftagu
An AI-powered clinical interview and triage system that conducts bilingual Urdu/English voice interviews, generates SOAP medical reports, and provides a dedicated doctor review portal.
Tech Stack
Next.js
React
TypeScript
TailwindCSS
PostgreSQL
Prisma
LangChain
ElevenLabs
The Problem
- Clinics lacked a scalable system to conduct structured patient interviews before physician appointments.
- Bilingual Urdu/English patients had no AI-powered tool that understood both languages for medical queries.
- Doctors had no automated way to receive pre-structured patient information ahead of consultations.
- Existing health chatbots could not generate standardised SOAP medical reports or flag emergency red flags.
- Language barriers and limited clinical hours created significant gaps in timely patient-physician communication.
Gallery
Our Solution
- Architected a full-stack clinical interview system using Next.js, LangGraph, and Groq (LLaMA 3.3 70B) for AI orchestration.
- Built voice-based bilingual interviews using ElevenLabs, automatically transcribing and structuring patient responses.
- Implemented Pinecone-based RAG for medical knowledge retrieval to support evidence-based response generation.
- Developed automated SOAP report generation from interview transcripts and emergency red-flag triage detection.
- Created a dedicated doctor portal for reviewing AI-generated reports, approvals, and prescription workflows.
Impact
Multi-lingualhealth AI platform
Enabled thousands of users to access reliable health consultations in their native language, reducing unnecessary clinic visits.
Ready to build something similar?
Let's discuss your project and see how we can help.