Ayusutra - Ayurvedic Wellness Platform
Full-stack wellness platform bridging Ayurvedic care and modern technology with AI therapy recommendations, clinic booking, and role-based dashboards
Timeline
Ongoing
Role
Full Stack & AI Developer
Team
Solo Project
Status
In-progressTechnology Stack
Key Challenges
- Designing scalable role-based access for users, clinics, and admins in one platform
- Serving real-time AI therapy recommendations through a FastAPI microservice
- Building an asynchronous ML pipeline for tongue analysis during onboarding without blocking UX
- Integrating secure phone OTP authentication and payment flows end-to-end
Key Learnings
- Architecting Next.js 14 full-stack apps with API-driven, multi-role product surfaces
- Deploying TensorFlow/Keras models behind FastAPI for production inference
- Using BullMQ and Redis for background ML jobs and responsive onboarding flows
- Shipping clinic dashboards, admin tooling, and user booking experiences in one codebase
Ayusutra - Ayurvedic Wellness Platform
Overview
Ayusutra is a next-generation platform that bridges ancient Ayurvedic wisdom and modern digital convenience. The mission is to make holistic wellness accessible, personalized, and engaging — with AI-powered therapy recommendations, clinic discovery, secure bookings, and dashboards for users, clinics, and admins.
Built as a scalable full-stack product with Next.js 14 on the frontend, FastAPI for AI and backend services, PostgreSQL for data, and TensorFlow/Keras for therapy recommendation models.
Live: ayusutra-v2.vercel.app · GitHub: Tusharkanta407/Ayusutra-v2
Problem Statement
Ayurvedic care is often fragmented — patients struggle to find trusted clinics, understand suitable therapies, and track personalized wellness journeys. Clinics lack unified tooling for appointments, patient context, and performance insights. Ayusutra centralizes discovery, booking, AI-guided recommendations, and operations in one modern platform.
System Architecture
User / Clinic / Admin (Next.js) → API Layer → PostgreSQL
→ FastAPI ML Service → TensorFlow/Keras Model
Onboarding ML Jobs → BullMQ + Redis
Auth → Firebase Phone OTP
Media → Cloudinary
Key Features
AI-Powered Therapy Recommendations
- Personalized Ayurvedic therapy suggestions based on health profile and preferences
- Custom TensorFlow/Keras model with label encoding and heuristic rules
- Served in real time via FastAPI microservice
Clinic Discovery & Booking
- Find top-rated Ayurvedic clinics and book appointments seamlessly
- Integrated payment gateway for secure therapy bookings
- User dashboard for bookings, therapy history, and recommendations
Clinic & Admin Operations
- Clinic dashboard: manage appointments, patient details, and clinic performance
- Admin panel: manage clinics, users, and platform content
- Role-based access across user, clinic, and admin surfaces
Authentication & UX
- Secure Firebase Auth with phone OTP login and registration
- Clean, responsive UI with Next.js and Tailwind CSS
- Mobile-friendly flows across onboarding and booking
Async ML Pipeline
- BullMQ + Redis queue for tongue-analysis ML during onboarding
- Improves responsiveness by offloading heavy inference from the main request path
Tech Stack
| Layer | Technology | |-------|------------| | Frontend | Next.js 14, React, TypeScript, Tailwind CSS | | Backend | FastAPI (Python) | | AI/ML | TensorFlow / Keras, custom Ayurvedic heuristic model | | Database | PostgreSQL | | Cache / Jobs | Redis, BullMQ | | Auth | Firebase Auth (Phone OTP) | | Media | Cloudinary | | Deploy | Vercel (frontend), FastAPI service |
Repository Structure
app/ → Next.js routes and pages
components/ → UI components
contexts/ → React context providers
fast-api/ → ML model and inference service
hooks/ → Custom React hooks
lib/ → Shared utilities
public/ → Static assets
styles/ → Global styles
types/ → TypeScript types
workers/ → Background job workers
What I Built
- Scalable full-stack wellness platform with role-based access for users, clinics, and admins
- AI-powered therapy recommendation system using TensorFlow/Keras, served via FastAPI
- Asynchronous processing pipeline with BullMQ and Redis for ML-based tongue analysis during onboarding
- Real-time admin and clinic dashboards for operational visibility
- Secure authentication, booking flows, and payment integration
Impact
- Unified Ayurvedic wellness experience from discovery to booking to personalized care
- Faster onboarding through background ML processing instead of blocking the UI
- Production-ready architecture separating web app, API, and ML inference layers
- Open, maintainable codebase structured for continued feature growth
Links
- Live App: https://ayusutra-v2.vercel.app/
- Repository: https://github.com/Tusharkanta407/Ayusutra-v2