### Answer Based on the API documentation, when creating a test case definition via the API, the JSON format should include the following structure: [(1) ]() [(2) ]()

json
{
  "name": "<string>",
  "response_engine": {
    "type": "retell-llm",
    "llm_id": "<string>",
    "version": 0
  },
  "user_prompt": "<string>",
  "metrics": [
    "<string>"
  ],
  "dynamic_variables": {},
  "tool_mocks": [
    {
      "tool_name": "<string>",
      "input_match_rule": {
        "type": "any"
      },
      "output": "<string>",
      "result": true
    }
  ],
  "llm_model": "gpt-4.1"
}

Key fields: - **name**: Name of the test case definition - **response_engine**: Response engine configuration (Retell LLM or custom LLM) - **user_prompt**: The user prompt to simulate in the test case - **metrics**: Array of metric names to evaluate - **dynamic_variables**: Dynamic variables to inject into the response engine - **tool_mocks**: Mock tool calls for testing - **llm_model**: LLM model to use for simulation (options include gpt-4.1, gpt-4.1-mini, gpt-4.1-nano, gpt-5, claude-4.5-sonnet, gemini-2.5-flash, etc.) However, I don't have specific documentation about uploading JSON files directly through the dashboard interface. The available documentation covers the API endpoints for creating and updating test case definitions programmatically. [(1) ]() [(2) ]() [(3) ]()

​ To learn more, see the sources I considered: [[1] Create Test Case Definition](https://docs.retellai.com/api-references/create-test-case-definition) [[2] Update Test Case Definition](https://docs.retellai.com/api-references/update-test-case-definition) [[3] List Test Case Definitions](https://docs.retellai.com/api-references/list-test-case-definitions) ​

Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.

### Answer Based on the available documentation, the JSON format I shared applies to test case definitions in general, but I need to clarify the response engine configuration for conversation flow specifically. [(1) ]() For conversation flow test cases, the

response_engine

field should use this structure:

json
{
  "type": "conversation-flow",
  "conversation_flow_id": "<string>",
  "version": 0
}

So the complete JSON format for a conversation flow simulation test case would be:

json
{
  "name": "<string>",
  "response_engine": {
    "type": "conversation-flow",
    "conversation_flow_id": "<string>",
    "version": 0
  },
  "user_prompt": "<string>",
  "metrics": [
    "<string>"
  ],
  "dynamic_variables": {},
  "tool_mocks": [
    {
      "tool_name": "<string>",
      "input_match_rule": {
        "type": "any"
      },
      "output": "<string>",
      "result": true
    }
  ],
  "llm_model": "gpt-4.1"
}

The key difference is the

response_engine.type

should be

"conversation-flow"

instead of

"retell-llm"

, and you'll need to provide a

conversation_flow_id

instead of an

llm_id

. However, I don't have specific documentation about uploading JSON files directly through the dashboard interface for test cases. The available documentation covers the API endpoints for managing test case definitions programmatically. [(2) ]() [(1) ]() [(3) ]()

Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.

​ To learn more, see the sources I considered: [[1] List Test Case Definitions](https://docs.retellai.com/api-references/list-test-case-definitions) [[2] Update Test Case Definition](https://docs.retellai.com/api-references/update-test-case-definition) [[3] Get Test Case Definition](https://docs.retellai.com/api-references/get-test-case-definition) ​