Welcome to HealthAI, an all-in-one health platform where you can track personal health, manage medications, and instantly screen for brain tumors from MRI scans using advanced AI.
HealthAI is a modern health companion—log your health metrics, medications, and treatments in one place. Plus, harness the power of AI with TumorSense to analyze MRI scans for brain tumors with just a few taps.
Main Features:
- MedLog: Track general health data, symptoms, medications, and treatments
- TumorSense: Upload and analyze MRI scans for brain tumor detection using AI
- Secure Cloud Storage: All health information is safely stored with Firebase
- Cross-Platform: Available on mobile (Thunkable) and web (React + Vite)
- Log and monitor symptoms, vitals, and general health data
- Record and manage your medications: name, dose, frequency
- Track ongoing and past treatments
- Securely save your personal health data in the cloud
- Sign in easily with email or Google (Firebase Auth)
- Instantly detect brain tumors from uploaded MRI images
- Powered by a convolutional neural network (TensorFlow + Teachable Machine, TFLite)
- Works seamlessly within HealthAI and as a standalone web app (included in this repo)
| Component | Description | Technology |
|---|---|---|
| HealthAI App | Full-featured health platform (med log + AI) | Thunkable (Mobile), React + Vite (Web) |
| TumorSense (Web) | Standalone brain tumor detection web app | React + Vite, hosted on Vercel |
| Backend API | AI prediction endpoint | Flask (Python), hosted on Render |
| AI Model | TumorSense AI (CNN for MRI scans) | TensorFlow, Teachable Machine (TFLite) |
| Image Management | MRI image upload and storage | Cloudinary |
| User Data & Auth | Secure data storage and authentication | Firebase Firestore + Auth |
- Architecture: Convolutional Neural Network (CNN)
- Framework: TensorFlow + Teachable Machine (TFLite)
- Input: 224 x 224 px grayscale MRI images
- Speed: ~0.3 seconds inference on low-resource cloud
- Output: Tumor class and confidence score (see API example below)
Send a JSON payload with the MRI image URL for brain tumor classification.
Request Example: { "image_url": "https://link-to-mri-image.jpg" }
text
Response Example: { "class": "Healthy", "confidence": 0.9989 }
text
- Upload MRI images directly from your phone
- Images are uploaded securely via Cloudinary
- Results delivered instantly from the Flask API
- All health data (med log, symptoms, medications) managed in-app
- Analyze MRI scans and get instant results
- Direct interface to the backend AI prediction service
- Try it now: https://health-ai-seven.vercel.app/
For detailed explanations, model training code, deployment (Render) instructions, and additional metrics related to the TumorSense web app, please see:
📎 Download TumorSense Web App Documentation (PDF)
The model was trained and evaluated with brain MRI datasets for reliable accuracy and fast inference time.
- User health data is encrypted and securely stored in Firebase Firestore
- Authentication is provided via Firebase Authentication (email & Google sign-in)
- MRI images are stored on Cloudinary with secure URLs
-
Mobile App Demo Video:
Watch here -
TumorSense Web App:
https://health-ai-seven.vercel.app/ -
TumorSense API Endpoint:
https://api-6903.onrender.com/predict
- Goutham Krishna D
- Govind Krishna D
Students, Indian School Al Seeb, Muscat, Oman
This repository includes:
- Full HealthAI mobile/web app (MedLog + TumorSense)
- Standalone TumorSense web app (React + Vite)
- Flask AI backend server
- Machine learning model files and training assets
Thank you for choosing HealthAI — your companion for smarter, connected healthcare and instant AI diagnostics.