How to Use ChatAnthropic with Langchain: Simple Guide

To use ChatAnthropic in Langchain, import it from langchain.chat_models and create an instance with your API key. Then call its generate method with a list of messages to get a chat response.

Syntax

The basic syntax to use ChatAnthropic involves importing the class, creating an instance with your API key, and calling the generate method with chat messages.

• Import: Bring in ChatAnthropic from langchain.chat_models.
• Instantiate: Provide your Anthropic API key to create the chat model.
• Generate: Pass a list of messages (with roles and content) to get the model's reply.

python

from langchain.chat_models import ChatAnthropic

chat = ChatAnthropic(anthropic_api_key="your_api_key")

response = chat.generate(messages=[
    {"role": "user", "content": "Hello!"}
])

print(response.generations[0].message.content)

Example

This example shows how to send a simple greeting message to ChatAnthropic and print the response text.

python

from langchain.chat_models import ChatAnthropic

# Replace with your actual Anthropic API key
chat = ChatAnthropic(anthropic_api_key="sk-xxxx")

messages = [
    {"role": "user", "content": "Hi there! How are you?"}
]

response = chat.generate(messages=messages)
print(response.generations[0].message.content)

Output

Hello! I'm doing well, thank you for asking. How can I assist you today?

Common Pitfalls

• Not providing a valid anthropic_api_key will cause authentication errors.
• Messages must be a list of dictionaries with role and content keys; otherwise, the call fails.
• Using roles other than user, assistant, or system may lead to unexpected behavior.
• Forgetting to access response.generations[0].message.content to get the text output.

python

from langchain.chat_models import ChatAnthropic

# Wrong: missing API key
# chat = ChatAnthropic()

# Wrong: messages not a list
# response = chat.generate(messages={"role": "user", "content": "Hello"})

# Right way:
chat = ChatAnthropic(anthropic_api_key="sk-xxxx")
response = chat.generate(messages=[{"role": "user", "content": "Hello"}])
print(response.generations[0].message.content)

Output

Hello! How can I help you today?

Quick Reference

Concept	Description
ChatAnthropic	Class to interact with Anthropic chat models in Langchain
anthropic_api_key	Your secret API key to authenticate requests
messages	List of dicts with roles ('user', 'assistant', 'system') and content strings
generate()	Method to send messages and get chat completions
response.generations[0].message.content	Access the text reply from the model

Key Takeaways

Always provide your Anthropic API key when creating ChatAnthropic instance.

Pass messages as a list of dictionaries with 'role' and 'content' keys.

Use the generate() method to get chat responses from the model.

Access the reply text via response.generations[0].message.content.

Avoid invalid roles and incorrect message formats to prevent errors.