Dynamically updating OpenAI Key

[blacksmithop]

We can dynamically update the temperature as so

model = ChatOpenAI(temperature=0).configurable_fields(
    temperature=ConfigurableField(
        id="llm_temperature",
        name="LLM Temperature",
        description="The temperature of the LLM",
    )
)
model.with_config(configurable={"llm_temperature": 0.9}).invoke("foo_prompt")

But the same cannot be done for api key

model = ChatOpenAI(openai_api_key="placeholderkey").configurable_fields(
    openai_api_key=ConfigurableField(
        id="openai_api_key",
        name="API Key",
        description="The API key used when making a call to OpenAI",
    )
)
model.with_config(configurable={"openai_api_key": somevalidapikey}).invoke("foo_prompt")

I realize this might be because of how secrets are configured

@property
    def lc_secrets(self) -> Dict[str, str]:
        return {"openai_api_key": "OPENAI_API_KEY"}

Can anyone explain if there is any workaround for this

[dosubot[bot]]

🤖

Based on the current design of the LangChain framework, the OpenAI API key is not designed to be dynamically updated in the same way as the temperature. The API key is set during the initialization of the BaseOpenAI class and is used to authenticate all subsequent requests to the OpenAI API.

The openai_api_key attribute is used in the validate_environment method to set the API key for the OpenAI client:

values["openai_api_key"] = get_from_dict_or_env(
    values, "openai_api_key", "OPENAI_API_KEY"
)

This means that when you update the openai_api_key attribute, the new API key will be used for subsequent calls to the OpenAI API.

However, you can manually update the openai_api_key attribute of an instance of the BaseOpenAI class as shown below:

# Assume `model` is an instance of the `BaseOpenAI` class
model.openai_api_key = 'new_api_key'

This will update the API key for the model instance. Please note that this will only affect the model instance and not any other instances of the BaseOpenAI class.

Please be aware that this is not the intended use of the openai_api_key attribute and may lead to unexpected behavior. It is recommended to set the API key during the initialization of the BaseOpenAI class and not change it afterwards.

For more details, you can refer to the source code of the BaseOpenAI class in the LangChain repository.

Sources

• libs/partners/openai/langchain_openai/llms/base.py

---

This response is meant to be useful and save you time. It is not meant to be a precise solution, but rather a starting point for your own research.

Help me be more useful! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

If you want to continue the conversation, start your reply with @dosu-bot.

[MahmutAtia]

did you find any solution?

[payalsaraljain]

Hi, is there any solution for it? I am facing the same issue