LifeLine

Inspiration

Emergency response systems face increasing pressure due to staff shortages, rising call volumes, and critical delays. Every second counts when saving lives, but human dispatchers are often overwhelmed. LifeLine was designed to bridge this gap—using AI to support dispatchers by gathering important information instantly and ensuring help arrives faster. Our mission is simple: No call goes unanswered, and no life is left waiting.

What it does

• AI-Powered Call Handling: When dispatchers are busy, LifeLine picks up calls, and collects important information like location, nature of the emergency, and any immediate dangers.
• Call Prioritization: The system sorts and prioritizes calls based on urgency, ensuring critical situations are addressed first.
• Emotion Detection: Using advanced NLP models, it identifies caller emotions (e.g., panic, fear, calm) to guide dispatchers in responding appropriately.
• Emergency Identification: AI classifies emergencies (medical, fire, crime) through real-time text analysis to assist in dispatching the correct units.
• Real-Time Mapping: Leaflet maps display live locations of emergencies and dispatched units, enabling precise tracking.
• Human in the Loop: AI performs the initial intake, but human dispatchers review all AI suggestions for accuracy and empathy.

How we built it

Frontend:

• HTML and Tailwind CSS for a responsive UI.
• Leaflet for interactive maps to display live locations of incidents and dispatched units. Backend:
• Python backend handles core logic, communication, and database operations.
• MongoDB stores and fetches call data, emergency details, and transcripts.
• Twilio facilitates voice calls and integrates speech-to-text conversion.
• Whisper for accurate transcription of emergency calls.
• OpenAI GPT API for cleaning and summarizing calls, extracting data
• Google Maps APIs for geocoding and Street View.

Challenges We Ran Into

One of the biggest challenges was managing latency issues. Integrating multiple APIs—such as Twilio for calls, Whisper for transcription, and Google Maps for tracking in real-time had noticeable delays. We optimized the process to reduce lag, ensuring critical data reached dispatchers without delays.

Accomplishments that We're Proud of

• Human-AI Collaboration: Designed a system where dispatchers stay in control, with AI offering real-time support.
• Optimized System for Low Latency: Overcame latency challenges by fine-tuning API integrations.

What We Learned

• Twilio and Whisper Integration: Learned integrating Twilio and Whisper for call handling and speech-to-text transcription.
• API Optimization with Google Maps: We learned skills to integrate Google Maps APIs effectively for geolocation services and Street View.
• MongoDB for Data Management: Learned how to use MongoDB to efficiently store and retrieve call data, and emergency details.

What's Next for LifeLine

• Predictive Analytics: Use historical data to predict peak call times and optimize dispatcher availability.
• AI-Driven Training Modules: Develop AI-powered training simulations based on real-world scenarios to train new dispatchers.

Built With

• ai
• flask
• javascript
• leaflet.js
• mongodb
• openai
• tailwind
• twilio