Gemini Alert

An AI-powered crisis de-escalation app that provides real-time guidance and connects users to help, ensuring safety and mental wellness during stressful situations.

Comment

Inspiration

We were inspired by the idea of integrating mental wellness and future cities to create a safer and more responsive environment for individuals facing stressful situations. As smart cities evolve, ensuring public safety and mental well-being should be a top priority. However, we realized that while large cities often have advanced emergency networks, smaller communities like Arlington, TX, lack similar resources to provide immediate assistance during crises. We wanted to bridge this gap by creating an AI-powered solution that not only offers personalized de-escalation strategies but also connects users to emergency contacts when needed. By combining mental health support with smart city responsiveness, Gemini Alert empowers individuals to manage crises effectively, making communities safer and more resilient.

We wanted to create an AI-powered solution that not only provides immediate, personalized de-escalation strategies but also connects users to emergency contacts if the situation escalates. Thus, Gemini Alert was born — a tool to empower individuals to manage crises effectively and safely.

What it does

Gemini Alert uses Google’s Gemini AI to:

  • Provide real-time de-escalation guidance for anxiety, conflict, and panic situations.
  • Allow users to trigger an emergency alert to their designated contacts if they need further assistance.
  • Offer step-by-step personalized support based on the situation’s severity.
  • Enable live streaming of AI responses for dynamic crisis situations.

How we built it

  • Backend: Flask API that integrates with Google’s Gemini API to generate real-time de-escalation responses.
  • Frontend: Built with Vue.js, ensuring a responsive and user-friendly interface.
  • Authentication: Firebase Auth to handle user authentication and token verification.
  • Database: Firestore to store user profiles and conversation history.
  • Cloud Functions: To trigger emergency contact notifications and manage real-time events.

Challenges we ran into

  • API Integration Issues: We faced initial difficulties in configuring the Gemini API and handling streaming responses efficiently.
  • Authentication Errors: Managing secure user authentication and token validation using Firebase required meticulous handling.
  • Circular Import Bug: Encountered a circular import issue in Firebase Admin that led to unexpected crashes, which was resolved after renaming conflicting files.

Accomplishments that we're proud of

  • Successfully integrating Gemini AI to provide real-time de-escalation strategies.
  • Implementing secure user authentication with Firebase and token verification.
  • Developing a prototype that streams AI-generated responses dynamically.
  • Creating a user-friendly interface with Vue.js that enhances the user experience.

What we learned

  • API Configuration & Security: Managing API keys securely and configuring Flask for real-time content generation.
  • Firebase Authentication & Firestore: Implementing secure login systems and handling user data efficiently.
  • Handling Edge Cases: Dealing with unexpected user inputs and ensuring the system provided appropriate fallback mechanisms.

What's next for Gemini Alert

  • Implementing real-time sentiment analysis to detect user distress and escalate assistance.
  • Expanding to multiple languages to make the app accessible to a global audience.
  • Integrating Apple Watch support to allow users to trigger emergency alerts and access de-escalation guidance directly from their wrist. This feature will enable users to discreetly send an alert using gestures like a double-tap or pressing the side button. Additionally, real-time heart rate monitoring can detect signs of distress or panic, prompting the app to provide immediate assistance automatically.
Share this project:

Updates