Langchain
- This is an opensource libary which
- abstracts the llm models
- base components for building
- rags
- agents (langgraph)
- tracing (observability) => Langsmith
-
Alternatives:
-
Langchain ecosystem

Hello – langchain
- Lets create a new project
mkdir hello-langgchain
cd hello-langgchain
uv init .
uv add langchain ipykernel
-
open vscode and
ctrl+shift+p=> Python Interpreter

-
create a new folder called as notebooks and create a file called as
hello-langchain.ipynband select kernel -
We need to understand two terms
- model: Name of the model which you are connecting
- model_provider: Provider of the model (GCP/AWS/Azure/….) if applicable.
-
to use a new model we need a package
langchain-openai, langchain-ollamaetc -
Refer Here for basic example.
-
Summary: Langchain abstracts models from your agents/rags.
