SilentSOS

Turning gestures into emergency alerts.

Comment

SilentSOS 🚨✋

Developer Week 2026 – Devpost Submission
When you can’t speak, SilentSOS does.

Inspiration

In emergency situations like kidnapping, domestic violence, or medical distress, victims often cannot shout or call for help. Inspired by real-world silent distress signals and the need for discreet emergency communication, we built SilentSOS — an AI-powered system that detects emergency hand gestures in real time and instantly alerts trusted contacts.

We wanted to create a solution where help can be requested without making a sound.

What it does

SilentSOS is a real-time emergency hand gesture detection and alert system.

  • Detects predefined emergency gestures (SOS, distress, kidnap alert)
  • Uses computer vision to track hand landmarks
  • Automatically sends SMS / WhatsApp alerts via Twilio
  • Shares live GPS location
  • Captures a screenshot as incident proof
  • Works in real-time using a webcam

It transforms silent gestures into immediate emergency notifications.

How we built it

SilentSOS is built using:

  • Python for backend logic
  • Mediapipe Hands for real-time hand landmark detection
  • OpenCV for webcam processing and image capture
  • Twilio API for sending SMS / WhatsApp alerts
  • Virtualenv for environment management

Pipeline:

  1. Webcam captures live video.
  2. Mediapipe extracts 21 hand landmarks.
  3. Gesture logic classifies specific emergency patterns.
  4. When detected:
    • Screenshot is captured
    • GPS location is fetched
    • Twilio sends alert message instantly

Challenges we ran into

  • Reducing false positives in gesture detection
  • Handling real-time performance without lag
  • Securely managing Twilio credentials using .env
  • Ensuring reliable message delivery
  • Designing gestures that are easy to perform but distinct enough for detection

Accomplishments that we're proud of

  • Successfully built a working real-time emergency detection system
  • Integrated gesture recognition + live alert automation
  • Achieved fast detection with minimal latency
  • Built a socially impactful AI solution
  • Designed a system that can scale to mobile and IoT devices

What we learned

  • Practical implementation of computer vision using Mediapipe
  • Real-time ML integration with external APIs
  • Handling environment variables securely
  • Building AI systems with real-world social impact
  • Importance of user-centric design in emergency tech

What's next for SilentSOS

  • Add voice-based emergency detection
  • Improve gesture classification using ML models
  • Deploy as mobile application (Flutter + TensorFlow Lite)
  • Smartwatch integration
  • Cloud dashboard for monitoring alerts
  • Expand gesture library for elderly & differently-abled users

🔹 Features

  • Detects emergency gestures using Mediapipe Hands
  • Sends SMS / WhatsApp alerts with live GPS location using Twilio
  • Real-time screenshot capture for incident proof
  • Lightweight and easy to deploy

🛠 Tech Stack

  • Python
  • Mediapipe
  • OpenCV
  • Twilio
  • Virtualenv

⚙️ Setup

git clone https://github.com/Harinath333/SilentSOS.git
cd help_me_alert_system
pip install -r requirements.txt

Built With

  • computer-vision
  • git
  • github
  • mediapipe-hands
  • opencv
  • python
  • real-time-hand-landmark-detection
  • twilio-api-(sms-&-whatsapp)
  • virtualenv
  • vs-code
Share this project:

Updates