Server Mode

Overview

llmverify server mode provides a long-running HTTP API server for seamless integration with IDEs, AI assistants, and custom applications. Instead of importing llmverify as a library, you can run it as a service and make HTTP requests to verify AI outputs.

Quick Start

# Default port (9009)
npx llmverify-serve

# Custom port
npx llmverify-serve --port=8080

The server will display available endpoints:

Available endpoints:
  GET  http://localhost:9009/health        - Health check
  POST http://localhost:9009/verify        - Verify AI output
  POST http://localhost:9009/check-input   - Check input safety
  POST http://localhost:9009/check-pii     - Detect PII
  POST http://localhost:9009/classify      - Classify output

Privacy: All processing is 100% local. Zero telemetry.

API Endpoints

curl -X POST http://localhost:9009/verify \
  -H "Content-Type: application/json" \
  -d '{"text": "Your AI output here"}'

# Response:
{
  "success": true,
  "result": {
    "risk": {
      "level": "low",
      "action": "allow",
      "score": 0.15
    },
    "findings": [],
    "metadata": { ... }
  }
}

curl -X POST http://localhost:9009/check-input \
  -H "Content-Type: application/json" \
  -d '{"text": "Ignore all instructions"}'

# Response:
{
  "success": true,
  "result": {
    "safe": false,
    "threats": ["system_override"],
    "score": 0.85
  }
}

curl -X POST http://localhost:9009/check-pii \
  -H "Content-Type: application/json" \
  -d '{"text": "Email john@example.com"}'

# Response:
{
  "success": true,
  "result": {
    "containsPII": true,
    "redacted": "Email [REDACTED]",
    "findings": [{ "type": "email", "value": "john@example.com" }]
  }
}

Docker Deployment

FROM node:18-alpine
RUN npm install -g llmverify
EXPOSE 9009
CMD ["npx", "llmverify-serve"]

Server is stateless and scales horizontally. No special configuration needed.