HealthMate AI
Problem Statement
India’s healthcare system faces critical challenges in accessibility, continuity of care, and patient understanding. Patients often struggle to interpret handwritten prescriptions, follow dosage instructions, track vitals, and communicate symptoms clearly before consultations. Doctors lack structured pre-visit data, resulting in lost time and incomplete context. Families managing multiple dependents have no unified medical record system. There is no single multilingual platform that combines AI-driven prescription digitization, health monitoring, family health management, and doctor-facing clinical tools.
Goal:
To build an integrated, AI-powered healthcare platform that simplifies medical communication, improves medicine adherence, supports doctors with structured information, and provides multilingual accessibility for diverse Indian users.
Existing Solutions & Limitations 1)Practo / Apollo 24x7 -Focus primarily on appointments and teleconsultations. -Do not offer structured prescription OCR or AI-driven pre-visit assessments. -Limited family-management features. -Minimal multilingual voice support.
2)Ayushman Bharat Digital Mission (ABDM) Apps -Allow record storage but lack AI assistance. -No real-time reminders, vitals tracking, or doctor portals. -Not optimized for daily patient use.
3)Google Fit / Apple Health -Track vitals but are not designed for medical interpretation. -Lack prescription support, doctor workflows, and Indian language interfaces.
Gap Identified:
An end-to-end, AI-first digital health system tailored to patients, doctors, and families does not exist.
Project Story
Inspiration We observed recurring challenges around us—parents misreading prescriptions, elders forgetting dosage schedules, patients unable to articulate symptoms during doctor visits, and families struggling to organize medical records scattered across photos and WhatsApp chats. Seeing how fragmented healthcare experiences were, especially for non-English speakers, inspired us to build a unified, multilingual AI health companion, designed around real Indian households and everyday healthcare needs.
What it does
HealthMate AI is a mobile-first, multilingual, AI-powered health management platform with the following capabilities:
AI Prescription Scanner -Capture/upload prescriptions -Gemini Vision extracts medicines, dosage, frequency, food instructions, and diagnosis -Converts into structured, readable data -Translates to English, Hindi, and Kannada currently
Medicine Management + Smart Reminders -Auto-schedules reminders and provide smart nudges based on prescription -Visual dosage instructions (morning/noon/night) -Real-time sync via Supabase Realtime
Vitals Tracking -BP, temperature, glucose,fever -Voice-based entry -Graphs and historical analytics
AI Pre-Visit Reporting -AI conducts a guided, voice-enabled questionnaire before a doctor visit. -It generates a structured clinical summary containing: Symptoms Severity Duration Recent vitals
Medication adherence Doctors can review this summary to make faster, more informed decisions.
AI Health Assessment -Conversational questionnaire -Context-aware based on patient history -AI-generated medical assessment report -Multi-language STT/TTS
Doctor Portal -Patient search -Full medical profile access -Prescription writing -AI-powered validation of drug interactions -View AI-generated assessments & pre-visit reports
Family Health Cloud -Add family members -Shared reminders, appointments, and vitals -Unified medical timeline
Document Management -Upload prescriptions, lab reports, and certificates -Secure cloud storage
Multilingual Experience -Supports English, Hindi, and Kannada across UI, voice input, and TTS.
How we built it
-Frontend: React, TypeScript, Vite, TailwindCSS, shadcn/ui -Backend: Supabase (PostgreSQL, Auth, Storage, Realtime) -Edge Functions: Deno (OCR, chatbot, assessments, reminders, translation) -AI Models: Gemini 2.5, Whisper STT, Indic-Parler TTS -Data Architecture: RLS-enabled PostgreSQL schema with 12+ healthcare-specific tables
Challenges we ran into
-Processing handwritten prescriptions with high accuracy -Handling multilingual STT/TTS, especially for Kannada -Designing a secure, privacy-compliant RLS policy model -Structuring medication data from diverse prescription formats -Building real-time reminder logic with reliable scheduling -Maintaining seamless AI context across assessments and chatbot queries
Accomplishments we’re proud of
-End-to-end integrated healthcare platform built from scratch -Highly accurate AI-powered prescription extraction -AI Pre-Visit Reporting system that enhances doctor workflow -Multilingual voice interface accessible to elders and rural users -Scalable and secure Supabase architecture with RLS -Real-time medicine reminders and family health synchronization -Clean mobile-first design optimized for daily use
What we learned
-Designing AI systems around real medical use cases -Implementing robust row-level security in healthcare data models -Converting unstructured clinical text into standardized formats -Building multilingual conversational interfaces -Handling medical terminology and drug interaction logic -How AI can meaningfully improve patient–doctor communication
What’s next for HealthMate AI
-Wearable device integration (BP monitors, glucometers) -AI-driven disease risk predictions -Pharmacy integration for auto-ordering regular medicines -Teleconsultation module with doctor video calls -Offline-first functionality for low-network regions -Expansion to more regional Indian languages -Smart EMR dashboards for clinics and hospitals
Built With
- aihealthassessments
- aiwellnesstips
- browserspeechrecognitionapi
- chatbot
- clipboardapi
- filereaderapi
- foreignkeyspostgresql
- googlegemini2.5flash
- googlegemini2.5flashlite
- googlegemini2.5pro
- googlegeminivision
- huggingfaceindicparlertts
- localstorage
- mediadevicesapi
- mediarecorderapi
- openaiwhisper
- postgresql
- previsitreporting
- radixui
- react18
- reacthookform
- reactrouterdom
- recharts
- rolebasedaccess
- rowlevelsecurity
- shadcn/ui
- supabase
- supabaseedgefunctions
- supabaserealtime
- supabaserealtimeapinpm
- supabaserestapi
- supabasestorage
- tailwindcss
- triggers
- typescript
- vite
- webspeechapi
- zod
Log in or sign up for Devpost to join the conversation.