What is LLM Wrapper?
The LLM Wrapper is a standardized interface that wraps around different LLM providers (OpenAI, Anthropic, Google, etc.). Instead of writing provider-specific code each time, you use one unified API. This makes switching between models easy.
LLM Wrapper is a layer that connects LangChain to an LLM (e.g., OpenAI API), so you can call the model using a unified interface.
What it does
- Wraps OpenAI API
- Standardizes model calling
- Makes switching models easier
How to use
from langchain_openai import ChatOpenAI
#You are no longer calling OpenAI directly.
llm = ChatOpenAI(
model="gpt-4o-mini"
)
response = llm.invoke("Explain RAG simply")
print(response.content)
or
llm = ChatOpenAI(
model="gpt-4o-mini",
temperature=0
)Previously, we directly call OpenAI AI (client = OpenAI(api_key=”your_api_key”) ), now we are no longer calling OpenAI directly, instead of
llm = ChatOpenAI(
model=”gpt-4o-mini”
)
for reviewing
| Directly call OpenAI AI (OpenAI SDK) | LangChain LLM Wrapper | |
| instantiate | client = OpenAI(api_key=”your_api_key”) | llm = ChatOpenAI( model=”gpt-4o-mini” ) |
| call | client.chat.completions.create(…) | llm.invoke(…) |