Safe-HearShot

"SafeShot — Real-time safety alerts from police scanners, right in your pocket."

Comment

SafeVoice Hearshot

Inspiration

The idea behind SafeVoice Hearshot emerged from two pressing issues:

  1. The fear of retaliation that silences victims of workplace harassment.
  2. The lack of real-time safety information during life-threatening incidents, like active shootings or emergencies.

One of our teammates once unknowingly entered a mall during a lockdown due to gunshots — despite 911 calls, no alerts surfaced on Citizen or Twitter. Similarly, many victims of harassment remain unheard due to broken reporting systems.

We wanted to empower people — whether they are victims of harassment or citizens facing safety threats — by giving them real-time, anonymous, and secure ways to report and receive crucial information.

Thus, SafeVoice Hearshot was born — blending anonymous reporting, real-time alerts, and support networks into a single platform.

How We Built It

We combined hardware and software innovations to create a powerful and affordable solution.

Tech Stack

  • Flutter — Cross-platform mobile app
  • Firebase — Authentication, backend, push notifications
  • Twilio — Anonymous messaging
  • Raspberry Pi + RTL-SDR — Hardware radio receiver
  • OpenAI Whisper — Real-time speech-to-text transcription
  • Claude LLM — Event location extraction
  • Google Maps API — Event mapping
  • Node.js — Messaging server
  • React Native (for early prototype)

Key Features

  • Anonymous Reporting — Users can report harassment safely and privately
  • Auto-generated Legal Complaints — Draft legal documents instantly
  • Real-time Incident Alerts — Stay informed about nearby dangers from police radio transcriptions
  • Access to Support Networks — Connect with NGOs, legal advisors, counselors
  • Peer Support Forums + Educational Content — Safe space to learn and support each other

What We Learned

  • How to build a real-time transcription pipeline using Whisper on noisy, low-quality radio signals
  • Integrating hardware radio receivers with software pipelines
  • Creating secure, anonymous communication channels
  • Designing an intuitive and user-centered interface using Figma
  • Balancing latency, accuracy, and reliability for safety-critical alerts

Challenges We Faced

  • Poor Radio Quality — Cleaning and processing low-quality, noisy police radio signals was a huge technical hurdle
  • Real-time Processing — Whisper AI wasn’t built for live transcription — we engineered our own custom layer on top of it
  • Cost Constraints — We hosted everything on our own self-hosted servers (GPUs and VPS) to minimize cloud expenses
  • Ensuring Anonymity & Security — Building robust anonymous reporting features while protecting user identities
  • Designing an Inclusive UX — Making sure both harassment victims and general users felt safe and empowered

Accomplishments We’re Proud Of

  • Our entire end-to-end pipeline works in real time — from radio reception to live notifications
  • Successfully created anonymous reporting and support tools for harassment victims
  • Blended hardware and software into a functional, low-cost, scalable system
  • Built a clean, modern mobile app and a functional Node.js backend
Share this project:

Updates