LSPD Interrogation MCP Server

A Model Context Protocol (MCP) based police interrogation simulation server powered by OpenAI.

📌 Key Features

MCP Integration:
- Built using Model Context Protocol SDK
- HTTP transport support
- Dynamic resource management (officer-profile, conduct-interrogation)
OpenAI Integration:
- Uses GPT-3.5-turbo model
- Generates dynamic interrogation strategies
- Simulates suspect responses
- Creates realistic dialogue flows
Core Components:
- Police officer profile management
- Smart interrogation mechanics
- Suspect behavior simulation
- Crime type and evidence integration

🚀 Installation

pnpm install
# Required environment variables
cp .env.example .env
# Start server
pnpm start

⚙️ Configuration

.env file:

OPENAI_API_KEY=your_api_key_here

Configurable parameters in config.ts:

AI model selection
Maximum token count
Temperature parameter (creativity level)

🌐 API Endpoints

Officer Profile

GET /profile/:badgeNumber

curl http://localhost:3000/profile/1234

Start Interrogation

POST /interrogations/{suspectId}

{
  "suspectName": "John Doe",
  "pressureLevel": 75,
  "crime": "Armed robbery",
  "evidence": ["Fingerprint", "Security camera footage"]
}

Suspect Response

POST /interrogations/{suspectId}/respond

{
  "suspectName": "John Doe",
  "officerStatement": "Your fingerprints were found at the crime scene!",
  "guilt": 85,
  "personality": "cowardly",
  "previousResponses": ["I'm innocent!"]
}

🔍 Example Usage

# Get officer profile
curl http://localhost:3000/profile/1234

# Start interrogation
curl -X POST http://localhost:3000/interrogations/suspect_01 \
  -H "Content-Type: application/json" \
  -d '{
    "suspectName": "John Doe",
    "pressureLevel": 80,
    "crime": "Drug trafficking",
    "evidence": ["Search records", "Confidential witness statement"]
  }'

✅ Data Validation

All endpoints include strong type checking and validation using Zod library:

Pressure Level: 0-100 (required)
Suspect Name: string format
Evidence: string array (optional)

🔒 Security

Sensitive data (OpenAI API key) managed through environment variables
HTTPS enforcement in production
Secure input handling with request validation

🤝 Contribution

Fork the repository
Create new branch (feat/my-feature or fix/issue-number)
Commit your changes
Push to the branch
Open a Pull Request

📜 License

Distributed under the MIT License.

Integrations