StreetGuardian AI is an AI-powered, voice-triggered emergency alert system designed for public safety and social good.
The system detects predefined emergency voice keywords and instantly triggers alerts with live visual evidence.
In emergency situations, victims may not be able to unlock their phone or make a call to seek help.
StreetGuardian AI enables hands-free emergency alerts using voice commands such as HELP or POLICE, allowing faster response in critical situations.
- Wake-word detection using Porcupine (Voice AI)
- Camera-based image capture using OpenCV
- Instant alerts sent to:
- Android Mobile Application (Firebase Cloud Messaging)
- Telegram (with captured snapshot)
- Local buzzer alert for nearby attention
- Fully standalone system with auto-start on boot
- Python
- OpenCV
- Porcupine (Wake-word Detection)
- Raspberry Pi
- Firebase Cloud Messaging (FCM)
- Android (Kotlin)
- Telegram Bot API
- Linux (systemd)
- Raspberry Pi
- USB Camera
- Microphone
- Buzzer
- Internet connectivity
A working demo of the physical prototype and alert system is available here:
🔗 **Demo Video: https://drive.google.com/file/d/1K81KrZylY7zTR3ZCY4YhclqvqL10HAIZ/view?usp=sharing 🔗 **Demo Video: https://www.linkedin.com/posts/nuthanaganti-chandrakanth-2149a02ba_streetguardianai-aiforsocialgood-raspberrypi-activity-7406334304026451968-6FCo?utm_source=social_share_send&utm_medium=member_desktop_web&rcm=ACoAAEytNzABNWfgLDCmz7v-N3ArLnC6vUP2P0k
Screenshots of the hardware setup and alert system are included in this repository.
🏅 Domain-Wise Best Team – AI for Social Good
HackVriksh 2025
This repository contains the prototype and demo implementation of StreetGuardian AI.
The final deployed Raspberry Pi system includes hardware-specific configurations and system-level services that are not included here.
This repository represents:
- Core system logic
- Voice-trigger workflow
- Camera capture
- Alert pipeline (Device → Cloud → Mobile)
Sensitive credentials (API keys, Firebase service accounts) are not included in this repository for security reasons.
This project is licensed under the MIT License.