AzureChatOpenAI returning None on errors, rather than raising an error

[jshutler]

Checked other resources

• I added a very descriptive title to this question.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.

Commit to Help

• I commit to help with one of those options 👆

Example Code

from langchain_openai import AzureChatOpenAI

llm = AzureChatOpenAI(
    deployment_name="gpt-4o",
    api_version="2024-08-06",
    api_key=<my-key>,
    azure_endpoint=<my_endpoint>,

)


response = llm.invoke("Hello, world")

print(response)
# prints none, and I get a log with a content moderation warning

Description

I'm currently building an agentic workflow that is using AzureChatOpenAI. I'm having an issue where one of my agents is building a prompt, giving it to another agent, but llm.invoke is returning None rather than raising an error.

I'm getting a log message saying its a moderation message saying I'm trying to jailbreak the system, which I am not trying to do.

What I want to do is get that log back as an error in an exception message that I can pass back to my prompt building agent, however as of now the invoke call just returns None.

Is there a way to set llm.invoke to return/raise this error, rather than returning None?

System Info

langchain==0.3.10
langchain-community==0.3.10
langchain-core==0.3.49
langchain-experimental==0.3.3
langchain-openai==0.3.11
langchain-postgres==0.0.13
langchain-text-splitters==0.3.2
langchainhub==0.1.14

[jshutler]

Ok, it turns out the issue was that I had a decorator on my call that was forcing a return to None. This is not an issue with langchain, my bad.