Usage

Bodyguard can be integrated in three ways:

1. HTTP Server

Deploy Bodyguard as a standalone service that other applications can query:
# GET request
curl "https://ai.civic.com/bodyguard/check?prompt=Your%20prompt%20here"

# POST request
curl -X POST https://ai.civic.com/bodyguard/check \
  -H "X-API-Key: your_api_key_here"
  -H "Content-Type: application/json" \
  -d '{"prompt":"Your prompt here"}'

2. Client Library

Integrate directly into your application:
import { check } from '@civic/bodyguard';

const { result, threatScore, findings } = await check('User prompt here', {
  threshold: 0.5 // Fail if threat score > 0.5
});

3. CLI Tool

bodyguard "Analyze this prompt for threats"

4. MCP Hooks

Bodyguard can be used as a middleware hook for MCP servers, analyzing prompts before they reach the LLM:
"mcpServers": {
  "my-mcp-server": {
    "command": "passthrough-mcp-server",
    "env": [
      "TARGET_SERVER_URL", "https://my-mcp-server",
      "HOOKS", "https://ai.civic.com/bodyguard/hook?threshold=0.7"
    ]
  }
}
For more details on the passthrough proxy, see the Pass-through Proxy documentation.

Docker Deployment

Docker images are available for easy deployment. Contact us to get access to the Docker images on AWS ECR.

Resources

For more information and access, please contact us.