LangChain

Architecture Overview

LangChain uses a modular architecture with core abstractions: Models (chat models, LLMs, embeddings), Prompts (templates, few-shot selectors), Output Parsers (JSON, Pydantic, etc.), Retrievers (vector stores, keyword search), and the LCEL (LangChain Expression Language) for composing components into chains. The Runnables protocol provides a unified interface for all components with built-in streaming, batching, and async support.

When to Use LangChain

• LLM-powered applications with rich integrations
• Chain-based workflows combining multiple LLM calls
• Tool and API integration via a standardized interface
• RAG applications with diverse retriever backends
• Prototyping with rapid model and tool swapping

Strengths & Weaknesses

Quick Start

from langchain_openai import ChatOpenAI
from langchain_core.prompts import ChatPromptTemplate
from langchain_core.output_parsers import StrOutputParser

# Create a simple chain with LCEL
prompt = ChatPromptTemplate.from_messages([
    ("system", "You are a helpful assistant that translates {input_language} to {output_language}."),
    ("human", "{input}"),
])

model = ChatOpenAI(model="gpt-4o")
output_parser = StrOutputParser()

# Compose with LCEL pipe operator
chain = prompt | model | output_parser

result = chain.invoke({
    "input_language": "English",
    "output_language": "French",
    "input": "Hello, how are you?",
})
print(result)

Features at a Glance

Notable Users

Resources