What this does

LangChain Expression Language (LCEL) provides a declarative way to compose chains using the | pipe operator. Each component (prompt, model, output parser) is a Runnable that can be chained together.

Steps

Set up the LLM and prompt. Create a ChatOllama instance and a prompt template.

from langchain_ollama import ChatOllama
from langchain.prompts import ChatPromptTemplate
from langchain.schema.output_parser import StrOutputParser

llm = ChatOllama(model="llama3.2", temperature=0.7)

prompt = ChatPromptTemplate.from_template(
    "You are a helpful assistant. Answer the question: {question}"
)

Compose the chain with LCEL. Use the pipe operator to connect prompt, model, and parser.

chain = prompt | llm | StrOutputParser()

Invoke the chain. Call invoke with input matching the prompt variables.

result = chain.invoke({"question": "What is the capital of France?"})
print(result)
# Output: The capital of France is Paris.

Add a RunnablePassthrough for debugging. Inject intermediate values without disrupting the chain.

from langchain.schema.runnable import RunnablePassthrough

chain = (
    RunnablePassthrough.assign(question=lambda x: x["question"])
    | prompt
    | llm
    | StrOutputParser()
)

Batch multiple inputs. LCEL supports batching natively.

results = chain.batch([
    {"question": "What is 2+2?"},
    {"question": "What is the capital of Japan?"}
])
for r in results:
    print(r)

Verification

python -c "
from langchain_ollama import ChatOllama
from langchain.prompts import ChatPromptTemplate
from langchain.schema.output_parser import StrOutputParser

llm = ChatOllama(model='llama3.2')
prompt = ChatPromptTemplate.from_template('Say: {word}')
chain = prompt | llm | StrOutputParser()
print(chain.invoke({'word': 'hello'}))
# Expected: hello
"

Common failures

Missing output parser. Forgetting StrOutputParser() returns an AIMessage object instead of a string.
Prompt variable mismatch. Template expects {question} but invoke passes {query}. Keys must match exactly.
Ollama not running. Chain hangs or raises ConnectionError. Verify with ollama list.
Version mismatch - The installed package or runtime differs from the command shown; check the version first and rerun the smallest verification command.
Local environment drift - Another service, virtual environment, model, or path is being used; print the active binary path and configuration before changing the guide steps.

Related guides

How to Compose Multiple Chains with Pipe Operator
How to Build Custom LLM Chain with Prompt Templates