Is there a way to instantiate an AzureChatOpenAI object that routes requests to a custom API which implements all relevant endpoints from OpenAI?

[Dennis-Gruendel-EAG]

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

const model = new AzureChatOpenAI({
    configuration: {
        baseURL: 'https://<CUSTOM_ENDPOINT>.azurewebsites.net',
        defaultHeaders: {
            "Authorization": `Bearer ${token}`
        },
    },
});

Description

I have a custom API that mimicks the chat/completions endpoint from OpenAI, but also does some necessary authentication which is why I also need to provide the Bearer token in the request header. As I am using the model for agentic workflows with several tools, I would like to keep using the AzureChatOpenAI class. Is it possible to set it up in a way where it only needs the URL of my backend API and the header, and it would call my backend API just like it would call the Azure OpenAI endpoint?

If I try to instantiate it like in my code example, I get:

Error: Azure OpenAI API key or Token Provider not found

And if I provide dummy values for azureOpenAIApiKey, azureOpenAIApiInstanceName, azureOpenAIApiDeploymentName, azureOpenAIApiVersion, my custom API does not register a call and I will get a connection timeout after more than a minute.

System Info

@langchain/openai@>=0.1.0 <0.6.0", "@langchain/openai@^0.5.5; @langchain/openai@~0.0.14
MacOS
Typescript 5.8.2