Digital Public Infrastructure (DPI) for Inclusive Healthcare. Bridging India's rural health gap through a multilingual, voice-first Digital Vaidya and a real-time Public Health Sentinel dashboard.
AyushDhara AI addresses the Trust Deficit and Accessibility Gap in Indian healthcare. With 70% of India's population in rural areas and practitioners concentrated in urban centers, our solution provides a "Health Guardian" in every pocket, grounded in verified AYUSH scriptures.
- Affordability: Scalable to ₹0.50 - 1.50 per consultation via AWS Serverless.
- Inclusion: Voice-first interaction in 5+ Indic languages (Hindi, Tamil, Telugu, Bengali, English).
- Public Impact: Real-time outbreak detection 2 weeks faster than traditional systems.
AyushDhara AI utilizes a highly scalable, serverless event-driven architecture designed for the "Next Billion Users."
- AI/LLM: Amazon Bedrock (Claude 3.5 Sonnet) using Retrieval-Augmented Generation (RAG).
- Voice/NLP: Amazon Transcribe (Indic accents) & Amazon Polly (Natural Indian voices).
- Data Layer: Amazon DynamoDB (Single-Table Design) & OpenSearch Serverless (Vector Store).
- Compute: AWS Lambda (Node.js 20.x) & Amazon API Gateway.
- Security: AWS KMS (PII Encryption) & Bedrock Guardrails.
An AI-driven interview to determine a user's Vata/Pitta/Kapha constitution, delivering hyper-personalized preventative care plans based on lifestyle data.
Powered by Amazon Bedrock, ensuring health guidance is strictly retrieved from verified AYUSH datasets to eliminate AI hallucinations.
A public health surveillance engine that anonymizes and aggregates community symptom data to visualize disease hotspots for government officials.
To ensure "Direct Selection" in the hackathon, we have prioritized ethics and safety:
- Emergency Hand-Off: A dedicated
lib/safety.tslayer intercepts "Red Flag" symptoms (e.g., chest pain, breathing difficulty) to immediately direct users to 108 Emergency Services. - DPDP Act 2023: Implements strict PII anonymization before data enters the Sentinel surveillance loop to protect citizen privacy.
- ABDM Ready: Data structures utilize FHIR R4 standards and SNOMED CT codes for seamless integration with India's national health mission.
- Node.js 20+
- AWS CLI configured for
ap-south-1(Mumbai) - Amazon Bedrock Model Access enabled for Claude 3.5 Sonnet
- Clone & Install:
git clone [https://github.com/niyatijain/ayushdhara-ai.git](https://github.com/niyatijain/ayushdhara-ai.git) cd ayushdhara-ai npm install