Jarvis Sentinel

About Jarvis Sent

Inspiration

In emergencies, the biggest problem isn’t the lack of help — it’s the delay in reaching it.
People often struggle to unlock their phones, make calls, or explain their situation under stress.

This project was inspired by a simple question:
“What if help could be triggered instantly — just by speaking?”

Jarvis Sentinel was built to reduce response time in critical moments and act when the user cannot.

-What It Does

Jarvis Sentinel is an AI-powered, voice-first emergency assistant.

With a simple wake command — “Hey Jarvis” — it:

• Detects the type of emergency (medical, fire, or security)
• Responds in real time with context-aware guidance
• Sends alerts with location to trusted contacts via WhatsApp
• Activates silent mode in dangerous situations (e.g., police emergency)

This ensures safety without escalating risk.

• How We Built It

The system combines:

• Frontend: HTML, CSS (Glassmorphism UI), JavaScript
  
• Backend: Node.js + Express
  
• AI Engine: Local LLM (Ollama - LLaMA 3.1)
  
• Voice System: Web Speech API (Speech Recognition + TTS)
  
• Emergency Logic: Custom multi-layer classifier (medical/fire/police)
  
• Integration: WhatsApp deep links for real-time alerting
  

The architecture is designed to be lightweight, fast, and responsive — ensuring minimal delay during emergencies.

• Key Features

• Voice-first interaction (hands-free operation)

• Multi-category emergency detection

• Silent alert system for high-risk situations

• Live location sharing (simulated for demo)

• Multi-contact emergency network (1–5 contacts)

• Secure onboarding with OTP-based authentication (simulated)

• What I Learned

• Designing for real-world stress scenarios requires simplicity and speed

• AI is not just about intelligence — it’s about context and human behavior

• UX decisions can directly impact safety and usability

• Building systems that feel real doesn’t always require full-scale infrastructure

• Challenges Faced

• Designing a reliable emergency detection system without false triggers

• Implementing silent mode without breaking user interaction flow

• Creating a realistic authentication system without backend services

• Balancing speed vs intelligence in AI responses

• Ensuring the system feels calm and human-like, not robotic

• Impact

Jarvis Sentinel is not just a project — it’s a step toward AI-assisted emergency response systems.

By reducing the time between danger and action,
it has the potential to save lives when every second counts.

Built With

• css-(glassmorphism-ui)
• express.js
• html
• javascript
• llama
• node.js
• ollama