Skip to content

High-stakes 112 dispatch simulator built with Murf Falcon. Features a "Panic Engine" that dynamically escalates voice emotion based on user response time. <500ms latency.

Notifications You must be signed in to change notification settings

Dupahar/crisis-call-murf-agent

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

🚨 CrisisCall: AI-Powered Stress Inoculation Platform

"We don't just simulate calls. We measure Grace Under Pressure."

CrisisCall UI

👨‍💼 Operator Dashboard (Active Call)

Operator Dashboard


� The Problem: The "Silence" Gap

High-stakes verbal skills are currently taught via:

  1. Passive Theory: Reading manuals (Zero emotional preparation).
  2. Human Roleplay: Hiring actors is expensive, unscalable, and inconsistent.

The Gap: There is no scalable way to teach people how to think clearly under auditory pressure.

⚡ The Solution: Dynamic Escalation Engine

CrisisCall is a browser-based training platform that uses a proprietary Dynamic Escalation Engine to monitor your response time and voice patterns in real-time.

  • If you are calm and fast (<2s): The AI de-escalates and provides critical information.
  • If you hesitate (>3s): The AI panics, screams, and increases the difficulty.
  • The Result: Personalized, high-fidelity stress training that scales infinitely.

🏆 Why Murf Falcon? (The "Psychological Realism" Layer)

We didn't just use Murf Falcon for speed; we used it for Psychological Realism.

In a crisis, silence creates panic. Standard TTS engines introduce a 2-second "robotic silence" that immediately breaks the immersion of a simulation. Murf Falcon's streaming capability eliminates this gap, achieving sub-500ms latency. This makes the simulation indistinguishable from a real life-or-death call.

🛠️ Tech Stack

Component Technology Role
Voice Generation Murf Falcon API Ultra-low latency Streaming TTS
Speech Recognition Deepgram Nova-2 Real-time ASR (Speech-to-Text)
Intelligence Google Gemini 2.5 Flash Low-latency Conversational AI
Frontend Next.js 14 React Framework & UI
Styling TailwindCSS Professional, dark-mode aesthetics
State Management React Hooks Custom "Panic" Logic & Audio Visualization

🚀 Setup & Installation

  1. Clone & Install:

    git clone https://github.com/Dupahar/crisis-call-murf-agent.git
    cd crisis-call-murf-agent
    npm install
  2. Environment Variables: Create a .env.local file in the root directory:

    MURF_API_KEY=your_murf_key
    DEEPGRAM_API_KEY=your_deepgram_key
    GEMINI_API_KEY=your_gemini_key
  3. Run Development Server:

    npm run dev

    Open http://localhost:3000 in your browser.


🎥 Simulation Walkthrough

Watch the Demo

To test the "Panic Mode" yourself:

  1. Click "Connect to Emergency Channel".
  2. Say: "112, what is your emergency?"
  3. Test 1 (Fast Response): Reply immediately.
    • Result: Anjali calms down and provides details.
  4. Test 2 (Hesitation): Wait 10 seconds in silence.
    • Result: Anjali screams, begs for help, and the UI pulses RED.

👥 Use Cases

  • B2G (Government): 112 Dispatch Training, Hostage Negotiation.
  • B2B (Corporate): PR Crisis Management, Cybersecurity Incident Response.
  • Healthcare: ER Triage and Telehealth Diagnostics.

License

MIT

About

High-stakes 112 dispatch simulator built with Murf Falcon. Features a "Panic Engine" that dynamically escalates voice emotion based on user response time. <500ms latency.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published