Radiance AI

Radiance AI — Your intelligent health companion for fast, trusted, and research-backed diagnosis.

Comment

Inspiration

Healthcare is something we all need—but it’s often hard to access, slow, expensive, or even intimidating. I’ve personally experienced moments where I felt sick or had persistent symptoms but hesitated to go to a doctor. Sometimes it was because of time or cost, but honestly, it was also because I felt nervous asking questions, or just too shy to talk about certain issues face-to-face. There were even times I googled symptoms late at night, unsure whether it was something serious or just nothing.

"I had questions about my health but was too afraid or embarrassed to ask. I wished there was someone—or something—I could trust to just listen and guide me."

That’s when it hit me—what if there was a way to get fast, reliable, and private health guidance right from your phone? Something intelligent enough to feel like talking to a real doctor, but without the fear or friction?

I've always wanted to build something meaningful in the healthcare space, but I wasn’t sure where to start. When I saw this hackathon and the power of the Perplexity Sonar API, it clicked: this is the perfect moment to begin. Radiance AI was born out of that spark—to make healthcare more accessible, intelligent, and comfortable for people like me, and millions of others who just want clear answers, fast.

What it does

Radiance AI is an advanced AI-powered health diagnosis platform that leverages cutting-edge artificial intelligence to provide users with preliminary health assessments. The application uses the Perplexity Sonar API to analyze user-reported symptoms and medical data, offering detailed diagnostic insights, treatment recommendations, and medical guidance.

Our innovative Chain Diagnosis System provides a comprehensive health assessment through multiple specialized AI roles, each powered by specific Perplexity Sonar models:

  • Medical Analyst (sonar-deep-research): Analyzes medical reports and images
  • General Physician (sonar-pro): Provides initial assessment and diagnosis
  • Specialist Doctor (sonar-reasoning-pro): Offers specialized medical insights
  • Pathologist, Nutritionist, Pharmacist (sonar-pro): Provide domain-specific recommendations
  • Follow-up Specialist (sonar-pro): Creates personalized follow-up plans
  • Radiance AI Summarizer (sonar-pro): Compiles all insights into a comprehensive report

How we built it

We built Radiance AI using a modern tech stack centered around the powerful Perplexity Sonar API:

  1. Frontend: Next.js 15 with React 19, TypeScript, and TailwindCSS for a responsive and accessible UI
  2. Backend: Serverless architecture with Next.js API routes and Supabase for authentication and database
  3. AI Integration: Perplexity Sonar API with specialized models for different medical roles:
    • Implemented a reusable API component with standardized payload structure
    • Created specialized system prompts for each AI role in the diagnosis chain
    • Integrated image analysis capabilities for medical reports and scans
  4. Data Storage: Supabase PostgreSQL for user profiles, medical history, and diagnosis results
  5. PWA Support: Progressive Web App capabilities for cross-platform accessibility

The core of our application is the Chain Diagnosis System, which orchestrates a sequence of specialized AI roles to provide comprehensive health insights. Each role uses a specific Perplexity Sonar model optimized for its function, with structured JSON outputs for consistent data handling.

Challenges we ran into

  1. AI Model Selection: Finding the right Perplexity Sonar models for each specialized medical role required extensive testing and prompt engineering
  2. Medical Image Analysis: Integrating image analysis capabilities with the Perplexity API required custom formatting and handling
  3. Streaming Responses: Implementing real-time streaming responses from the Perplexity API while maintaining UI responsiveness
  4. JSON Structure Consistency: Ensuring consistent JSON structure from different AI models for reliable data processing
  5. Mobile Responsiveness: Creating an intuitive flow visualization for the Chain Diagnosis progress indicator on mobile devices
  6. Authentication Integration: Seamlessly integrating Supabase authentication with Next.js while maintaining session persistence

Accomplishments that we're proud of

  1. Multi-Agent Chain Diagnosis: Successfully implementing a chain of specialized AI roles using different Perplexity Sonar models for comprehensive health analysis
  2. Medical Image Analysis: Enabling the analysis of medical reports and images through the Perplexity API
  3. Structured JSON Responses: Creating a reliable system for structured data extraction from AI responses
  4. Intuitive Progress Visualization: Developing an engaging visual representation of the diagnosis process with cosmic-themed animations
  5. Comprehensive Health Reports: Generating detailed health insight reports with actionable recommendations
  6. Responsive Design: Creating a fully responsive application that works seamlessly across devices

What we learned

  1. Perplexity Sonar API Capabilities: We gained deep insights into the capabilities of different Perplexity Sonar models and how to optimize prompts for medical analysis
  2. Prompt Engineering: We learned advanced techniques for crafting effective system prompts that guide AI models to produce structured, reliable outputs
  3. AI Chain Architecture: We developed expertise in designing and implementing multi-agent AI systems where each agent has specialized knowledge and responsibilities
  4. Image Analysis Integration: We discovered effective methods for integrating image analysis capabilities with text-based AI models

What's next for Radiance AI

Radiance AI is evolving into a ready-to-launch AI-powered healthcare SaaS platform. We’re preparing for a beta release that will allow early adopters to experience intelligent, research-backed health diagnostics through a user-friendly, cross-platform interface.

If we win this hackathon, it will unlock major opportunities—from validation and visibility to vital funding that we’ll use to:

  1. Expand infrastructure and real-time AI capabilities
  2. Enhance medical image analysis and wearable data integration
  3. Accelerate HIPAA-compliant deployment and offline-first functionality

We're also looking for a strategic partnership with Perplexity to co-develop healthcare-focused AI tools, optimize model performance, and scale Radiance AI globally with their cutting-edge Sonar technology.

With a strong foundation, a working prototype, and a bold vision, Radiance AI is on the path to revolutionize digital health diagnostics and bring accessible, AI-powered healthcare to everyone.

Built With

Share this project:

Updates