Articles → LANGCHAIN → Message Classes In Langchain

Message Classes In Langchain

This article explains the message classes in Langchain.

Message Classes

In langchain, we have the following message classes: -

HumanMessage
SystemMessage
AIMessage

Humanmessage

Represents the message sent by the user.

Systemmessage

Represents system-level instructions that guide the model’s behavior.

Aimessage

Represents a response generated by the AI assistant.

Example

from langchain_openai import ChatOpenAIfrom langchain_core.messages import HumanMessage, SystemMessage, AIMessage# Create modelllm = ChatOpenAI(    model="gpt-4o-mini",    temperature=0)# Conversation messagesmessages = [    SystemMessage(content="You are a helpful coding assistant."),    HumanMessage(content="What is Python?"),    AIMessage(content="Python is a programming language."),    HumanMessage(content="Who created it?")]# Send conversation to modelresponse = llm.invoke(messages)print(response.content)

Output

How Does The Flow Work?

SystemMessage(content="You are a helpful coding assistant.")

These are the instructions to the model. The model uses this as the global rule for the entire conversation.

HumanMessage(content="What is Python?")

User asks the first question.

 AIMessage(content="Python is a programming language.")

Represents a previous assistant response that you manually add to the conversation history. It helps the model understand past replies so it can maintain context. In short, you are simulating a chat where the assistant has already answered before.

HumanMessage(content="Who created it?")

The user asks another question. The model will finally respond because it has the context from the previous messages.

Posted By -	Karan Gupta

Posted On -	Thursday, May 14, 2026

Query/Feedback

Your Email Id		**

Subject		*

Query/Feedback	Characters remaining 250	**