POST /mltk/aicommander_metadata

Description. Live-test a provider configuration without saving it. The handler issues a probe request against the provider using the supplied or previously-saved credentials and reports latency, HTTP status, and a snippet of the response.

Body (JSON). Same shape as POST /mltk/aicommander (above). The handler uses servicesettings.access_token.value; if absent it pulls from the saved secret keyed by connection_name + id.

Response (200 on success).

{
  "message": "Successfully connected to OpenAI. Model 'gpt-4o-mini' is accessible.",
  "status": "success",
  "provider": "openai",
  "model": "gpt-4o-mini",
  "response_content": "Hello! How can I assist you today?",
  "response_time": 1.23,
  "http_status_code": 200
}

Response (failure).

{
  "message": "Authentication failed. Please check your OpenAI API key.",
  "status": "fail",
  "provider": "openai",
  "model": "gpt-4o-mini",
  "http_status_code": 401,
  "response_time": 0.45
}

Example.

curl -sk -u "$SPLUNK_USER:$SPLUNK_PASSWORD" -X POST 
  -H "Content-Type: application/json" 
  --data @- "$SPLUNK_HOST/servicesNS/nobody/Splunk_ML_Toolkit/mltk/aicommander_metadata?output_mode=json" <<'JSON'
{
  "service": "OpenAI",
  "model":   "gpt-4o-mini",
  "servicesettings": {
    "endpoint":     { "value": "https://api.openai.com/v1/chat/completions" },
    "access_token": { "value": "sk-..." }
  },
  "modelsettings": {
    "max_tokens":     { "value": 50 },
    "connection_name":{ "value": "default" }
  }
}
JSON

Prerequisites by provider.

Provider	Required fields
OpenAI	endpoint, access_token
Azure OpenAI	endpoint, access_token, azure_api_version
Anthropic	endpoint, access_token
Bedrock	region, aws_access_key_id, aws_access_token, role_arn
Splunk Hosted Models	Valid SCS token, configured in mlspl.conf
Ollama	endpoint (no token typically)

Source. bin/rest_handlers/aicommander.py, bin/rest_handlers/aicommander_metadata.py, bin/ai_commander/ai_commander_util.py, bin/ai_commander/llm_factory.py.