I'm trying to call the

start

method with an assistant config object. This was working just a week ago, but now I'm getting an error that

assistant

is a null object -- which is definitely not the case.

Traceback (most recent call last):
  File "/Users/andrewchang/Documents/acalabs/ai/product/elixir/elixir-py/examples/apps/vapi-chatbot/scripts/call_vapi_agent.py", line 43, in <module>
    vapi.start(assistant=assistant)
  File "/Users/andrewchang/Documents/acalabs/ai/product/elixir/elixir-py/examples/apps/vapi-chatbot/.venv/lib/python3.12/site-packages/vapi_python/vapi_python.py", line 37, in start
    call_id, web_call_url = create_web_call(
                            ^^^^^^^^^^^^^^^^
  File "/Users/andrewchang/Documents/acalabs/ai/product/elixir/elixir-py/examples/apps/vapi-chatbot/.venv/lib/python3.12/site-packages/vapi_python/vapi_python.py", line 22, in create_web_call
    raise Exception(f"Error: {data['message']}")
Exception: Error: Couldn't Get Call Resources. Error: Cannot read properties of null (reading 'firstMessage')

https://cdn.discordapp.com/attachments/1263279035392458752/1263279035589328997/call_vapi_agent.py?ex=6699a7c8&is=66985648&hm=e1a81bedea2a604b95eca83d78767aefccfccac3506ac749727672a440f40cd6&

@achang97 It seems like there might be an issue with the assistant configuration object you're passing to the

start

method. The error message indicates that within the

create_web_call

function, it's trying to access a property

firstMessage

on a null object, which suggests that the assistant object might not be properly initialized or there's a mismatch in the expected structure. To resolve this, please double-check the assistant configuration object you're providing to ensure it matches the expected format and contains all the necessary properties, including

firstMessage

. If you're certain that the assistant object is correct and you're still encountering this issue, it might be a problem with the VAPI itself or a change in the API that hasn't been accounted for in your code. If you need further assistance with the assistant configuration or if the problem persists, please provide the configuration object you're using (ensuring to redact any sensitive information), and I'll be happy to help you troubleshoot the issue further. If the issue is beyond the scope of what can be resolved here, you may need to reach out to the VAPI support team for more in-depth assistance.

Can you reach out to the team

@achang97 for custom llm was it working before with same api endpoints?

Yep, it was working before

@achang97 please check dm.