What is LangGraph?
LangGraph is an orchestration framework for complex agentic systems and is more low-level and controllable than LangChain agents. On the other hand, LangChain provides a standard interface to interact with models and other components, useful for straight-forward chains and retrieval flows. LangGraph leverages graph theory in order to orchestrate agents and forward instructions between them.
Why use LangGraph?
Other agentic frameworks can work for simple, generic tasks but fall short for complex tasks bespoke to a company’s needs. LangGraph provides a more expressive framework to handle companies’ unique tasks without restricting users to a single black-box cognitive architecture.
Basic Principles
States
States represent the current data available at any point in the workflow. It is a shared memory bank for all nodes in the graph. It is used to hold intermediate values as your graph executes node-by-node. It is the first thing to define when constructing a graph in langgraph.
Nodes
Nodes execute based on the instructions given to them when they are instantiated.
Linking Nodes and Functions
You can bind the nodes to different functions using llm.bindtools([*function name*]), where the llm object is an instantiation of the LLM API. For example:
from langchain_openai import ChatOpenAI
llm = ChatOpenAI(model="gpt-4o")
def multiply(a: int, b: int) -> int:
"""Multiply a and b.
Args:
a: first int
b: second int
"""
return a * b
llm_with_tools = llm.bind_tools([multiply])
Edges
Edges determine the flow of the application.