-
-
Healix AI – AI-powered health intelligence using Gemini 3
-
Users enter symptoms to receive AI-driven health insights and predictions.
-
Symptom-based disease prediction with confidence scoring using ML models.
-
Gemini 3–powered conversational assistant providing explainable, context-aware health guidance.
Inspiration
The inspiration behind Healix AI came from a simple but serious problem: when people feel unwell, their first step is often random symptom searches on the internet, which leads to confusion, fear, and misinformation. As a student exploring AI for social impact, I wanted to build something that feels like a calm, intelligent first point of guidance—not a replacement for doctors, but a responsible assistant that helps users understand their condition better before seeking professional care.
With the power of Gemini 3, I saw an opportunity to move beyond static symptom checkers and create a system that can reason, explain, and guide holistically.
What it does
Healix AI is an AI-powered health intelligence assistant that helps users make sense of their symptoms and recovery steps:
Symptom-Based Disease Prediction: Users enter symptoms manually or conversationally, and Healix AI predicts the most likely condition with confidence scores.
360° Recovery Roadmap: Instead of just naming a disease, the system generates structured guidance across:
- Precautions
- Common Medications
- Diet Recommendations
- Safe Workouts / Physical Activity
Conversational Health Assistant: Powered by Gemini 3 Pro, users can ask follow-up questions and understand why certain precautions or habits are recommended.
How we built it
This project was built using a vibe coding approach with Google AI Studio, focusing on rapid iteration and reasoning-first design:
AI Core: Used Gemini 3 Pro as the main reasoning engine for medical explanations, follow-ups, and contextual understanding.
Prompt Engineering: Carefully designed system prompts to ensure medical grounding, safe responses, and explainable outputs.
Backend: Python + Flask to connect symptom data, ML predictions, and Gemini-generated explanations.
Frontend: A clean, minimal UI focused on clarity and reduced anxiety, optimized for both desktop and mobile users.
Challenges we ran into
Medical Safety & Hallucinations: Health is a sensitive domain. We mitigated this by using strict system instructions and limiting speculative outputs.
Confidence vs Fear Balance: The challenge was to inform users without alarming them—Gemini 3’s reasoning helped frame responses calmly.
UI Logic: Preventing greetings or casual chat from triggering medical predictions required intent-aware conversation handling.
Accomplishments that we’re proud of
Built a complete AI health assistant using Gemini 3 within a hackathon timeframe.
Successfully combined ML predictions + generative reasoning into a single workflow.
Delivered a project focused on real-world social impact, not just technical novelty.
What we learned
Reasoning matters more than raw prediction in healthcare AI.
Gemini 3 excels at contextual explanations, making AI outputs more trustworthy.
Vibe coding with Google AI Studio drastically reduces development friction and boosts creativity.
What’s next for Healix AI
Medical Report & Image Analysis using Gemini 3 vision capabilities.
Doctor-Shareable Health Summary (PDF) for faster consultations.
Personalized Health Tracking with future wearable data integration.
Language Localization to make healthcare AI accessible in regional languages.
Log in or sign up for Devpost to join the conversation.