Chat Models and the Pipe (LCEL) Chain

The pipe operator builds a sequence

Every LangChain component that can be invoked is a Runnable with the same methods: invoke, batch, stream and async versions. The | operator (LangChain Expression Language, LCEL) joins runnables into a RunnableSequence: the output of each step is the input of the next. A typical chain is prompt | model | parser. Because the interface is the same, you can swap the real model for a fake model in tests, add retries, or stream, without rewriting the chain. A chat model takes messages and returns an AIMessage; StrOutputParser extracts the text.

prompt | model | parser, run

I ran this offline in a Python virtual environment with langchain-core 1.6.6, langchain-text-splitters 1.1.2 and llama-index-core 0.14.25. No API key or network call is needed because a fake model or a toy embedding stands in for the real one. A FakeListChatModel returns a canned reply, so the chain runs offline. The result is a RunnableSequence whose steps are the prompt, the model and the parser.

from langchain_core.prompts import ChatPromptTemplate
from langchain_core.output_parsers import StrOutputParser
from langchain_core.language_models.fake_chat_models import FakeListChatModel

prompt = ChatPromptTemplate.from_template("Translate to French: {text}")
model = FakeListChatModel(responses=["Bonjour le monde"])   # stands in for a real chat model
chain = prompt | model | StrOutputParser()

print(chain.invoke({"text": "Hello world"}))
print(type(chain).__name__)
print([type(step).__name__ for step in chain.steps] if hasattr(chain, "steps") else "")

Output:

Bonjour le monde
RunnableSequence
['ChatPromptTemplate', 'FakeListChatModel', 'StrOutputParser']

Swapping in a real model (illustrative)

Only the model line changes; the rest of the chain is identical. Requires the provider package and an API key in the environment; not run here.

from langchain_openai import ChatOpenAI          # pip install langchain-openai
from langchain_core.prompts import ChatPromptTemplate
from langchain_core.output_parsers import StrOutputParser

prompt = ChatPromptTemplate.from_template("Translate to French: {text}")
model = ChatOpenAI(model="<model-name>", temperature=0)   # reads OPENAI_API_KEY
chain = prompt | model | StrOutputParser()
print(chain.invoke({"text": "Hello world"}))

Quick check: What does the | operator do in LCEL?

• Starts a server
• Performs a bitwise OR on strings
• Deletes a step
• Connects runnables so each output feeds the next step