Dec 14, 2025

Sarthi Ai

healthcare ai assistant hackathon sarthi agentic

Sarthi AI - Intelligent Healthcare Companion

Project Overview Sarthi AI is an AI-powered healthcare "Middle Layer" designed to bridge the gap between patient confusion and professional medical care. By acting as an intelligent triage assistant, it simplifies the healthcare journey—helping patients understand their symptoms, find the right specialist instantly, and maintain a unified medical history.

The Problem: Friction in Modern Healthcare

The Dangers of Self-Diagnosis: Patients often rely on unverified internet searches for non-severe symptoms, leading to misdiagnosis, anxiety, and dangerous self-medication.

 The "Distance" Barrier: Long queues and logistical hurdles make hospitals feel inaccessible, causing patients to delay care until minor issues become emergencies.

Misrouting & Errors: Patients frequently visit the wrong specialist due to a lack of knowledge, driving up costs and delaying correct treatment.

 Fragmented Medical History: Doctors often lack a complete view of a patient's health because records are scattered across different clinics and physical files.

The Solution: A Data-Driven Ecosystem

 AI-Equipped Context-Aware Diagnosis: Sarthi uses Google's Gemini API to analyze symptoms against the patient's specific medical history, offering personalized, safer guidance than generic web searches.

 Precision Specialist Routing: The system identifies the exact medical need and directs the patient to the correct specialist (e.g., Cardiologist vs. General Physician) immediately.

Longitudinal Health Record: A "living" health timeline that updates in real-time. Every chat, symptom, and appointment is logged, giving doctors a complete view of the patient's health before the visit.

 Instant SOS Protocol: A One-Tap SOS feature that calls an ambulance, shares live GPS location with contacts, and books an emergency appointment instantly.

Tech Stack

Frontend Architecture

  • React (Vite): Utilized for its high-performance component-based architecture and rapid development build speeds.
  • TypeScript: Implemented for type safety, reducing runtime errors and ensuring robust, maintainable code quality.
  • Context API: Used for efficient global state management across the application without third-party bloat.

AI & Intelligence

  • Google Gemini API: The core intelligence engine powering the symptom triage and medical context analysis.
  • Custom Prompt Engineering: Tailored system instructions to ensure medical safety, empathy, and accurate specialist routing.

User Interface & Design

  • Tailwind CSS: Employed for a utility-first styling approach, enabling a fully responsive and modern design system.
  • Lucide React: A consistent and clean icon library to enhance visual communication and user experience.
  • Dark Mode Support: Integrated theme switching for accessibility and user comfort.

Development & Deployment

  • Vite: selected as the build tool for its lightning-fast HMR (Hot Module Replacement) and optimized production bundles.
  • ESLint: Ensures code consistency and enforces best practices.
  • Vercel / GitHub Pages: Scalable cloud hosting solutions for seamless deployment.

This build was uploaded as a hackathon project

Hackathon

HackSpace

View All Projects
Give a star to encourage!Discussion
Start a new conversation!
Login to join the discussion