----------------------------- **While I generate an answer, here are some sources that might be helpful**: * [Function Calling]() * [List Calls]() * [Server Events]() * [Tools Calling]() -----------------------------

### Answer The error

pipeline-error-azure-openai-llm-failed

may be related to the specific way properties are defined in your function call. Vapi supports custom functions with parameters, including arrays, but it's crucial to ensure the schema is properly defined and matches the expected format. Here's an example of how to define a custom function with an array parameter:

json
{
    "functions": [
        {
            "name": "exampleFunction",
            "description": "A function that takes an array of strings.",
            "parameters": {
                "type": "object",
                "properties": {
                    "arrayParameter": {
                        "type": "array",
                        "items": {
                            "type": "string"
                        }
                    }
                }
            }
        }
    ]
}

[(1)]() , [(2)]() Ensure that your function schema correctly specifies the type and structure of the array. Also, you may want to verify the request payload against the schema to ensure compliance. If the error persists, it may be related to a bug or an implementation detail specific to the Vapi pipeline. For further assistance, you might need to verify the correct implementation details or any updates from Vapi related to handling arrays.