Doctor Copilot

Overview

Doctor Copilot is an AI-powered assistant designed to support mental health professionals by providing valuable insights into a patient’s digital behavior — without the need for chatbot interaction or active check-ins.

It passively collects data such as:

• App usage and screen time
• Visited websites and their categories
• Notifications and mental-health-related keywords

This data is analyzed to generate summary reports, behavioral insights, and graphical charts, helping doctors focus on key behavioral indicators and patterns for more effective consultation and diagnosis.

🔒 Privacy-first: The system is designed with a strong emphasis on data confidentiality. While the current proof of concept performs analysis on a server (due to time constraints and integration complexity), our vision is to use llama.rn to run AI agents directly on the mobile device, ensuring all raw data stays local and only anonymized reports are sent to the doctor.

Features

• ✅ One-time setup by the patient with permission approvals
• ✅ A chatbot to help doctors gain deeper insights into patient reports for better diagnosis and treatment.
• ✅ Continuous passive data collection from the phone
• ✅ Categorization of websites and apps used
• ✅ Extraction of mental-health-related keywords from notifications
• ✅ Summary report with actionable key points
• ✅ Full report with behavior analysis
• ✅ Visual charts to support understanding of behavioral patterns

System Components

📱 Mobile App

A lightweight Android app installed on the patient's phone.

Responsibilities:

• Collects app usage time
• Tracks visited websites and classifies them
• Extracts notification keywords related to mental health

Example of data collected :

Tech Stack: Java, Android Studio

📸 Screenshots:

• 
• 

🎥 Demo Video:
Demo : Click Here

AI Backend (Proof of Concept)

A Flask server using AI agents built with CrewAI to analyze collected data and generate structured reports.

Responsibilities:

• Process raw data
• Generate summary and full reports
• Extract trends and key topics
• Prepare data for dashboard visualizations

Tech Stack:

• Python
• Flask
• CrewAI
• Gemini API

💻 Doctor Dashboard

A web dashboard for mental health specialists to manage patients, visualize behavioral data, and access generated reports.

Responsibilities:

• View patient summary and detailed reports
• Visualize behavioral trends with charts
• Manage patient profiles and appointment notes

Tech Stack:

• Next.js
• Tailwind CSS
• TypeScript

📸 Screenshots:

• 
• 

🎥 Demo Video:
Demo : Click Here

🔁 Workflow

graph LR
A[Patient's Phone] -->|Data Collection| B[Mobile App]
B -->|Sends Data| C[Flask Server]
C -->|AI Analysis via CrewAI| D[Report Generation]
D -->|Structured Report + Graphs| E[Doctor Dashboard]

Loading

🚀 Future Improvements

• Integrate llama.rn to run CrewAI agents directly on-device
• Encrypt local data and use secure enclave processing
• Implement real-time alerts for emergency behavior detection
• Add support for iOS via React Native

📁 Repository Structure

DoctorCopilot/
│
├— mobile-app/         # Android application
├— backend/            # Flask + CrewAI processing
├— frontend/          # Next.js doctor interface
└— README.md

👥 Team Name: CPU Team

Made with ❤️ during AI wave– 2025.