VAPI•2y ago

Configuring an assistant in an assistant-request response

I see in the docs and also successfully tested this with the following response from my serverUrl:

{
    "assistant": {
        "firstMessage": "Hey Ryan, how are you?",
        "model": {
            "provider": "openai",
            "model": "gpt-3.5-turbo",
            "messages": [
                {
                    "role": "system",
                    "content": "You're Ryan's assistant..."
                }
            ]
        }
    }
}

{
    "assistant": {
        "firstMessage": "Hey Ryan, how are you?",
        "model": {
            "provider": "openai",
            "model": "gpt-3.5-turbo",
            "messages": [
                {
                    "role": "system",
                    "content": "You're Ryan's assistant..."
                }
            ]
        }
    }
}

{
    "assistant": {
        "firstMessage": "Hey Ryan, how are you?",
        "model": {
            "provider": "openai",
            "model": "gpt-3.5-turbo",
            "messages": [
                {
                    "role": "system",
                    "content": "You're Ryan's assistant..."
                }
            ]
        }
    }
}

{
    "assistant": {
        "firstMessage": "Hey Ryan, how are you?",
        "model": {
            "provider": "openai",
            "model": "gpt-3.5-turbo",
            "messages": [
                {
                    "role": "system",
                    "content": "You're Ryan's assistant..."
                }
            ]
        }
    }
}

However, I'm also trying to add more configuration like functions, maxDuration, maxTokens, etc. that are available on the create assistant API. Now I'm getting an "assistant-request-returned-invalid-assistant" error but with no details about what is causing the problem. An example of my response is this:

{
    "assistant": {
        "firstMessage": "Hi!",
        "maxDurationSeconds": 89,
        "maxTokens": 250,
        "functions": [
            {
                "name": "get_weather",
                "async": false,
                "description": "Get the weather",
                "parameters": {
                    "type": "object",
                    "properties": {
                        "zip": {
                            "description": "This is the user provided zip code",
                            "type": "string"
                        }
                    }
                },
                "serverUrl": "https://serverurl.com"
            }
        ],
        "transcriber": {
            "provider": "deepgram",
            "model": "nova-2-phonecall",
            "language": "en-US"
        },
        "model": {
            "provider": "openai",
            "model": "gpt-4o",
            "messages": [
                {
                    "role": "system",
                    "content": "You are very helpful..."
                }
            ]
        }
    }
}

{
    "assistant": {
        "firstMessage": "Hi!",
        "maxDurationSeconds": 89,
        "maxTokens": 250,
        "functions": [
            {
                "name": "get_weather",
                "async": false,
                "description": "Get the weather",
                "parameters": {
                    "type": "object",
                    "properties": {
                        "zip": {
                            "description": "This is the user provided zip code",
                            "type": "string"
                        }
                    }
                },
                "serverUrl": "https://serverurl.com"
            }
        ],
        "transcriber": {
            "provider": "deepgram",
            "model": "nova-2-phonecall",
            "language": "en-US"
        },
        "model": {
            "provider": "openai",
            "model": "gpt-4o",
            "messages": [
                {
                    "role": "system",
                    "content": "You are very helpful..."
                }
            ]
        }
    }
}

{
    "assistant": {
        "firstMessage": "Hi!",
        "maxDurationSeconds": 89,
        "maxTokens": 250,
        "functions": [
            {
                "name": "get_weather",
                "async": false,
                "description": "Get the weather",
                "parameters": {
                    "type": "object",
                    "properties": {
                        "zip": {
                            "description": "This is the user provided zip code",
                            "type": "string"
                        }
                    }
                },
                "serverUrl": "https://serverurl.com"
            }
        ],
        "transcriber": {
            "provider": "deepgram",
            "model": "nova-2-phonecall",
            "language": "en-US"
        },
        "model": {
            "provider": "openai",
            "model": "gpt-4o",
            "messages": [
                {
                    "role": "system",
                    "content": "You are very helpful..."
                }
            ]
        }
    }
}

{
    "assistant": {
        "firstMessage": "Hi!",
        "maxDurationSeconds": 89,
        "maxTokens": 250,
        "functions": [
            {
                "name": "get_weather",
                "async": false,
                "description": "Get the weather",
                "parameters": {
                    "type": "object",
                    "properties": {
                        "zip": {
                            "description": "This is the user provided zip code",
                            "type": "string"
                        }
                    }
                },
                "serverUrl": "https://serverurl.com"
            }
        ],
        "transcriber": {
            "provider": "deepgram",
            "model": "nova-2-phonecall",
            "language": "en-US"
        },
        "model": {
            "provider": "openai",
            "model": "gpt-4o",
            "messages": [
                {
                    "role": "system",
                    "content": "You are very helpful..."
                }
            ]
        }
    }
}

Ask Inkeep•6/4/24, 1:34 PM

Hi @natanio, I'm an AI assistant for Vapi. I'll analyze available sources and get back to you shortly.

Ask Inkeep•6/4/24, 1:34 PM

-----------------------------
While I generate an answer, here are some sources that might be helpful:

Create Phone Call
Create Assistant
Server Events
List Calls
-----------------------------

Ask Inkeep•6/4/24, 1:34 PM

Answer

It seems like you are encountering issues when trying to add additional configurations to your assistant via the assistant-request from your

serverUrl

serverUrl

. The error you mentioned, "assistant-request-returned-invalid-assistant," typically indicates that the assistant configuration you're passing doesn't conform to the expected schema defined by Vapi.

Based on your configuration snippet, a few things need to be verified:

Parameter and Field Validation: Ensure all fields and parameters are named and used correctly as per Vapi's standards. You mentioned using
```
maxDurationSeconds
```
```
maxDurationSeconds
```
and maxTokensmaxTokens, which are not part of the assistant's settings retrieved or modified through the
```
serverUrl
```
```
serverUrl
```
response in the documentation.
JSON Structure: The JSON must be correctly structured and all required fields must be included. If the functionsfunctions array or any other components are not structured or named exactly as expected, that could trigger an error.

Since the detailed schema for what can be returned in an assistant-request isn't explicitly covered in the provided documents, I recommend ensuring that your implementation strictly adheres to what is defined in the Vapi documentation for creating or updating assistants. Here are some key aspects you can check:

**Transcriber and Model Compliance**: Ensure the `transcriber` and `model` configuration adhere strictly to the schemas provided in the documentation. For instance, the model type (`nova-2-phonecall` for `transcriber` or `gpt-4o` for `model`) should be validated against what's permissible in the API references.
**Function Configuration**: The structure of the `functions` array should precisely match the expected format. Make sure that all necessary sub-fields are provided and correctly formatted.

For now, you may need to limit adjustments through the

serverUrl

serverUrl

to parameters known to be suppo

Ask Inkeep•6/4/24, 1:34 PM

rted, and use the official API endpoints for any configurations not documented as part of the server response schema.

To resolve the issue, I recommend reviewing the API documentation related to creating and updating assistants to ensure your configurations are correct and supported. You can refer to the documents on creating assistants (1) and server events (2) for more detailed guidance.

Also, consider contacting Vapi support directly if the configuration aligns with the documentation but the problem persists.

Sources
[1] Create Assistant
[2] Server Events

AAsk Inkeep rted, and use the official API endpoints for any configurations not documented a...

Ask InkeepAPP•6/4/24, 1:39 PM

Got it. Tagging @Vapi Team for additional help.

AAsk Inkeep rted, and use the official API endpoints for any configurations not documented a...

Ask InkeepAPP•6/4/24, 2:30 PM

Got it. Tagging @Vapi Team for additional help.

living-lavender•6/5/24, 3:46 AM

There must be something wrong. Please check your JSON data at https://jsonlint.com/.

JSON Online Validator and Formatter - JSON Lint

JSONLint is the free online validator, json formatter, and json beautifier tool for JSON, a lightweight data-interchange format. You can format json, validate json, with a quick and easy copy+paste.

foolish-indigoOP•6/7/24, 3:03 AM

The JSON is valid

living-lavender•6/7/24, 9:30 AM

Can you send me the exact curl request? I will check it on my side.

Configuring an assistant in an assistant-request response

Answer

Similar Threads

Similar Threads

Answer

Similar Threads