Inspiration
We wanted to make healthcare more accessible and inclusive by simplifying how patients book and manage doctor appointments. Many patients, especially seniors or those with language barriers, find traditional systems difficult to use. Our inspiration was to create a conversational, voice-ready assistant that helps anyone interact with hospitals naturally — just by chatting.
What it does
Sunrise Medical AI Appointment Assistant allows users to:
- Chat with an AI assistant to book, cancel, or check appointments.
- View available doctors and their details in real time.
- Receive SMS and email notifications using AWS SNS.
- Use natural language (and future voice input) for accessibility.
It connects patients, doctors, and hospital systems through a single intuitive AI interface.
How we built it
- Frontend: Next.js + ShadCN UI for a responsive, accessible web interface.
- Backend: Flask REST API hosted on Render for serverless scalability.
- AI Layer: Amazon Bedrock (Claude 3.5 Sonnet, Haiku) for intent extraction and conversation understanding.
- Database: Amazon DynamoDB stores doctors, appointments, and patient data.
Notifications: AWS SNS for SMS/email alerts and reminders. Architecture: Fully cloud-based with Bedrock, DynamoDB, SNS,
Challenges we ran into
- Integrating multiple AWS services (Bedrock, SNS, DynamoDB) smoothly within Flask.
- Managing cross-origin requests between the Flask API and Next.js frontend.
- Handling structured AI responses consistently from Bedrock models.
- Ensuring accessibility and clean UI design under tight hackathon deadlines.
Accomplishments that we're proud of
What we learned
- How to orchestrate multiple AWS services (Bedrock, DynamoDB, SNS) in a real-world application.
- Best practices for AI-driven backend workflows and intent extraction.
- Deploying fullstack apps with Next.js + Flask and managing CORS, environment variables, and APIs efficiently.
- The value of teamwork and time management in rapid prototyping.
What's next for Sunrise Medical AI Appointment Assistant
- Add automated reminders via EventBridge + Lambda.
- Integrate Amazon Polly + Translate for multilingual voice support.
- Expand to handle medical report summaries and teleconsultation scheduling.
- Deploy a fully serverless version using AWS Lambda and API Gateway.
Built With
- amazon-bedrock
- amazon-dynamodb
- amazon-sns
- boto3
- flask
- next.js
- shadcn