Skip to main content
Sign up and grab your api keys at https://console.halios.ai/sign-up.

​
Authentication

Include your API key in the X-HALIOS-API-KEY header: 
curl -X POST "http://api.halios.ai/api/v3/agents/{agent_id}/evaluate" \
  -H "X-HALIOS-API-KEY: your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": "Hello, can you help me with something?"
      }
    ],
    "invocation_type": "request"
  }'

​
Core Endpoint

POST /api/v3/agents/{agent_id}/evaluate Evaluates content against configured guardrails for a specific agent. Messages need to be in OpenAI compatible message format. invocation_type accepts two values - request and response, indicating whether you are evaluating ai request or response.

​
Request Body

{
  "messages": [
    {
      "role": "user",
      "content": "Your message content here"
    }
  ],
  "invocation_type": "request"
}

​
Response

{
  "response_id": "12345-uuid-here",
  "guardrails_applied": 3,
  "guardrails_triggered": 0,
  "processing_time_ms": 145.2,
  "result": []
}

​
Integration Flow

  1. Before LLM Call: Send user input to evaluate endpoint
  2. Check Results: If any guardrail triggered, block the request
  3. Proceed: If no triggers, continue with LLM processing
  4. After LLM Call: Optionally evaluate LLM response
See the OpenAPI docs for complete endpoint details.
⌘I