SymptoWise - AI Symptom Checker

🧠 Inspiration

Many people experience discomfort or early symptoms but hesitate to seek help due to uncertainty, anxiety, or lack of clear guidance. We wanted to bridge that gap by creating a tool that not only predicts potential conditions based on symptoms but also empowers users with actionable advice and communication support — especially for those preparing for a doctor visit.

💡 What it does

SymptoWise is an AI-powered symptom checker that allows users to:

• Select symptoms via a searchable, multi-select interface
• Receive top 3 condition predictions with confidence levels
• View detailed condition descriptions and immediate precautions
• Chat with an integrated LLM assistant that helps phrase symptoms and recommend next steps

The platform combines predictive modeling with natural language guidance to improve user confidence and pre-consultation clarity.

🛠️ How we built it

• Frontend: React + TypeScript + TailwindCSS for a modular and responsive card-based UI
• Charts & Animations: Recharts for visualization, Framer Motion for UI transitions
• Symptom Input: react-select for searchable, tag-style symptom selection
• Backend: Python API (FastAPI) serving a trained classification model (joblib)
• LLM Assistant: Ollama running a local instance of LLaMA 3 for generating care guidance
• Markdown Rendering: react-markdown for formatting chatbot responses

🧩 Challenges we ran into

• Ensuring the ML model output was both accurate and user-friendly
• Handling local LLM responses with markdown formatting and scroll-safe UI
• Designing an interface that was intuitive, especially for non-technical or anxious users
• Efficiently passing state between components while keeping the UI reactive and modular

🏆 Accomplishments that we're proud of

• Building a fully working AI health assistant in just a few days
• Integrating a local LLM (Ollama) into the frontend for real-time care guidance
• Designing a smooth, mobile-friendly UI that feels polished and supportive
• Keeping both predictive accuracy and user trust central to our experience

📚 What we learned

• How to integrate Ollama locally into a React-based frontend
• Best practices for combining ML predictions with conversational UX
• Importance of thoughtful UX design in health tools — especially around clarity, tone, and accessibility
• Managing and visualizing confidence scores in an interpretable way

🚀 What's next for SymptoWise – AI Symptom Checker

• 🗣️ Add voice-to-text support for symptom input
• 🤖 Fine-tune the LLM to provide more medically-grounded advice
• 📱 Launch a PWA version for mobile access
• 📄 Export personalized care summaries as PDFs
• 🌐 Multilingual support for greater accessibility
• 🧑‍⚕️ Build an API for clinics to integrate SymptoWise into virtual triage flows

Built With

• axios
• csv
• fastapi
• framer-motion
• joblib
• llama-3
• node.js
• ollama
• pandas
• python
• react
• react-markdown
• react-select
• recharts
• tailwind-css
• typescript
• vite