Inspiration
As freshmen living far from home—navigating the streets of Los Angeles and exploring San Francisco—we often found ourselves wondering: "How safe am I right now?" Existing safety resources are scattered, inconsistent, and often fail to reflect real-time or hyperlocal risk factors like time of day or personal context.
We wanted to make it effortless to understand not only how safe a city is, but how safe your exact neighborhood or route might be—instantly and intelligently.
What It Does
SafetyBuddy Pro is your personal safety companion. Within seconds, it provides a clear, data-driven safety score for your current area or any destination you plan to visit.
The app automatically fetches your location and integrates multiple live data sources such as Civic Hub, Google Maps, and Apple Maps to assess:
- Recent crimes reported nearby
- Proximity to police stations
- Changes in risk level based on time of day
- Trends in local safety over the past few weeks
Using our Safety Metric Algorithm, this data is synthesized into a simple color-coded safety score—Safe, Moderate, or Dangerous—that’s glanceable at a moment’s notice.
SafetyBuddy also considers personal context. By optionally analyzing a quick selfie using Lava and OpenAI’s vision models, the app can understand visible belongings, perceived risk indicators, and demographic context like age and gender. This personal layer allows SafetyBuddy to generate tailored recommendations to help you stay safe in your environment.
If you’re heading somewhere new, you can search the destination to see an instant safety report and preview the risk level of each neighborhood along your route. In emergencies, SafetyBuddy provides SOS tools like direct 911 calling, an emergency whistle, and a flashing light alert to attract attention.
How We Built It
We designed SafetyBuddy Pro to feel both intuitive and powerful:
- Frontend built in Swift and SwiftUI for a clean, responsive iOS experience
- Backend powered by FastAPI, acting as a secure proxy between the app and cloud-based AI models
- Data integration through scraping Civic Hub, Apple Maps, and Google Maps using BeautifulSoup
- AI processing and recommendations handled via Claude through Lava APIs
This architecture allows fast, low-latency responses and dynamic insights tailored to the user’s context.
Challenges We Ran Into
We initially aimed to integrate live social media insights so users could see what people on the ground were saying about safety in real time, but API paywalls and access restrictions made that unfeasible for now. Balancing accuracy, speed, and privacy also required careful data design and filtering.
Accomplishments We’re Proud Of
- Developed a clean, minimalist UI that communicates complex data at a glance
- Built adaptable and efficient endpoints for near-real-time data retrieval
- Created a safety scoring algorithm that meaningfully blends location, time, and personal context
What We Learned
We learned how to integrate new APIs through Lava and FastAPI pipelines, efficiently scrape and normalize civic data across multiple sources, and design interfaces that balance aesthetics with clarity so even complex analytics remain accessible.
What’s Next for SafetyBuddy Pro
We plan to:
- Integrate social media APIs to surface real-time safety sentiment from locals
- Add an iOS Live Activity widget for instant SOS access from the lock screen
- Expand route safety visualization, showing risk levels by neighborhood and public transport line
- Enhance AI-driven personalization so safety insights adapt dynamically to surroundings and context
Summary
SafetyBuddy Pro is a smart, personal safety companion that turns scattered data into actionable, glanceable insights, helping you stay safe anytime and anywhere.
Built With
- beautiful-soup
- claude
- fastapi
- swift


Log in or sign up for Devpost to join the conversation.