SafeOStroll

Your ultimate safety companion. Stay safe, stay connected!

Comment

Inspiration

The idea for SafeOStroll came from those moments of unease many of us experience while walking alone at night. Imagine heading home after a long day or en route to meet friends. You pass through an area that feels unsafe, glancing over your shoulder, alert to every sound. That lingering sense of vulnerability is something most of us can relate to.

This inspired us to create SafeOStroll—a personal safety tool designed to be your constant companion. Whether navigating dimly lit streets or unfamiliar environments, SafeOStroll provides reassurance and peace of mind.

"As someone who often feels uneasy walking at night but has no choice but to do so after work or university, this app is a game changer for me. It gives me confidence knowing I have a companion to talk to—whether it's offering advice on safety or simply a comforting conversation." — Arina

Our goal is not just to enhance individual safety but to foster a greater sense of security in communities. By integrating emergency support, community alerts, and AI-driven emotional assistance, SafeOStroll aims to empower individuals and build safer environments worldwide.

What It Does

SafeOStroll is your ultimate safety companion, always by your side when you need it most. Whether you’re heading home late, passing through unfamiliar neighborhoods, or just feeling uneasy, SafeOStroll ensures you're never alone.

The app’s core feature is its AI-powered emotional assistant, providing comfort and guidance in stressful moments. Whether walking down an empty street or facing an uncertain situation, the AI offers reassurance, advice, and support—like having a trusted companion with you at all times.

This AI assistant engages in calming conversations and delivers personalized suggestions for handling anxiety or unsafe scenarios. It ensures that even when you're nervous, you're never left feeling isolated.

Behind the scenes, SafeOStroll's emergency features are always ready to activate. In a crisis, you can trigger an instant response—alerting both authorities and nearby users. Your location is shared, creating a safety network that mobilizes in real-time.

SafeOStroll goes beyond physical safety, addressing emotional well-being by combining AI-driven support with a community safety network, ensuring users feel connected and secure as they navigate public spaces.

Stay calm. Stay connected. Stay safe.

How We Built It

Building SafeOStroll involved combining advanced technologies with a user-centered approach, ensuring seamless functionality and safety:

AI-Powered Emotional Support

At the heart of SafeOStroll is our AI-powered assistant, developed using OpenAI’s GPT-4, tts1-hd, and whisper-1 API. This AI engages in calming conversations and offers actionable advice during stressful situations. The AI continuously learns from user interactions, improving its ability to provide personalized support. We also utilized WebSockets to ensure real-time communication between users and the AI assistant, creating a more interactive and responsive experience. The goal: ensure users never feel alone.

Real-Time Emergency Response

The app’s emergency features are built on a robust Django backend. With one tap, users can send distress signals to 911 and notify nearby users. Cloudflare ensures fast and secure transmission of real-time data, offering safety at your fingertips.

Mobile-Optimized Frontend

Using React, we built a mobile-first interface that delivers a seamless experience across devices. The app updates user coordinates every 10 seconds, providing real-time tracking in emergencies.

Design & Security

The user interface is both calming and intuitive, with gradient designs and hover effects creating a sense of reassurance. Data is secured with encryption, ensuring all user information stays private.

Challenges We Faced

Building SafeOStroll presented unique challenges that tested our technical and creative abilities.

1. AI Responsiveness

Creating an AI that felt natural while offering timely advice was a key challenge. We had to balance providing calming conversations with actionable suggestions, ensuring the AI felt supportive but not clinical. Also, making it so that the AI would have its own allocated memory was a challenge we had to overcome.

2. Real-Time Location Tracking

Implementing accurate real-time location tracking without draining users' battery required significant optimization. We needed to maintain frequent updates while minimizing energy consumption.

3. Data Privacy & Security

Handling sensitive user data, like locations and emergency signals, raised significant privacy concerns. We had to ensure all communications were encrypted while keeping the app responsive.

4. User-Friendly Design

Creating an intuitive, reassuring interface for users in distress was more challenging than expected. We had to ensure that the emergency and AI features were easy to access without overwhelming the user.

5. API Integration

Since this was our first time integrating the OpenAI API, figuring out how to get the different AI systems (TTS and STT) to interact with the user while still having an optimal response time was challenging. We also had to be careful with the training we gave the AI, as we didn't want the AI to act as a therapist but rather as a friend who can give you specialized advice.

Accomplishments We're Proud Of

We achieved several key milestones in developing SafeOStroll, each reflecting our dedication to creating a reliable and secure safety tool.

1. AI-Driven Emotional Support

Our AI assistant offers real-time emotional support using OpenAI’s GPT-4, tts1-hd, and whisper-1 API, providing calming conversations and personalized advice that adapts over time through integration.

2. Diversity

Another part of our project that we are proud of is that we can provide AI conversations in all current live languages, making it so that all users can communicate with the AI through their native tongue.

3. Seamless Real-Time Emergency Response

We developed a real-time alert system that connects users with emergency services and nearby SafeOStroll users. Powered by Django and Cloudflare, this system ensures distress signals are transmitted securely and swiftly.

4. Optimized Location Tracking

Our software updates user coordinates every 10 seconds to ensure accurate & precise location tracking without significant battery drain, improving emergency response accuracy.

5. User-Centered Design

Our mobile-first interface prioritizes ease of use, making it simple for users to send alerts, access the AI assistant, and navigate features during stressful moments.

6. Robust Data Privacy & Security

We ensured all user data is encrypted, providing a secure experience without compromising performance.

7. Secured Connection

We secured the connection between our React frontend and the insecure backend host through Cloudflare, enhancing the overall security of the application.

What We Learned

The SafeOStroll development journey taught us valuable lessons about technology, design, and user needs.

1. User-Centered Design

We learned the importance of constant iteration and feedback in creating a user-friendly interface, especially for users in distress.

2. AI Empathy

Designing an AI that provides emotional support without seeming robotic was challenging. We learned the importance of natural conversation flow and empathetic responses.

3. Security Is Essential

Handling sensitive user data highlighted the need for robust encryption and privacy protocols to maintain user trust and protect their information.

4. Optimizing Real-Time Systems

We gained insight into optimizing real-time systems, ensuring fast, reliable, and energy-efficient performance.

5. WebSockets

We learned that we can use WebSockets for communication between server and client for having stateful conversations.

What's Next for SafeOStroll

SafeOStroll’s journey is far from over, and we have exciting plans for the future.

1. Expanding AI Capabilities

We plan to further enhance the AI’s ability to provide tailored support, learning from user interactions to offer more personalized advice.

2. Health Data Tracking Integration

In the future, using technologies like Fitbit or Apple Watch, we can track the user's health data (such as BPM and stress detection) to have the AI more accurately assess the user's real-time situation.

Share this project:

Updates