Deploy on LangGraph Cloud

Author: JoonHo Kim
Peer Review :
Proofread : fastjw
This is a part of LangChain Open Tutorial

Overview

LangGraph Cloud is a cloud-based framework designed to simplify the development, deployment, and management of graph-based workflows for AI applications. It extends the functionality of LangGraph by providing a scalable, distributed, and user-friendly environment to build complex AI agents, workflows, and pipelines.

With LangGraph Cloud, you can:

Handle large workloads with horizontally-scaling servers, task queues, and built-in persistence.
Debug agent failure modes and quickly iterate in a visual playground-like studio.
Deploy in one-click and get integrated tracing & monitoring in LangSmith.

This tutorial will guide you through the key features and components of LangGraph Cloud, including:

Setting up LangGraph Cloud: How to create an account, configure your workspace, and deploy your first workflow.
Deploying workflows: Deploying workflows using LangGraph Cloud.
Using LangGraph Studio: How to connect to the Web UI LangGraph Studio and test the assistant.
Testing the API: How to send messages to the assistant using the Python SDK and verify the message data using Rest API.

By the end of this tutorial, you will be equipped with the knowledge to effectively utilize LangGraph Cloud for building and managing AI workflows in a scalable and efficient manner.

Now, let's dive in and explore how to boost performance with LangGraph Cloud! 🚀

References

Environment Setup

Set up the environment. You may refer to Environment Setup for more details.

[Note]

langchain-opentutorial is a package that provides a set of easy-to-use environment setup, useful functions and utilities for tutorials.
You can checkout the langchain-opentutorial for more details.

%%capture --no-stderr
%pip install -qU langchain-opentutorial

from dotenv import load_dotenv
from langchain_opentutorial import set_env

# Attempt to load environment variables from a .env file; if unsuccessful, set them manually.
if not load_dotenv():
    set_env(
        {
            "LANGCHAIN_TRACING_V2": "true",
            "LANGCHAIN_ENDPOINT": "https://api.smith.langchain.com",
            "LANGCHAIN_PROJECT": "LangGraph-Cloud",
        }
    )

# Install required packages
from langchain_opentutorial import package

package.install(
    [
        "langgraph-sdk",
    ],
    verbose=False,
    upgrade=False,
)

Prerequisites

Before we start, ensure we have the following:

Setting up a new repository on GitHub

To deploy a LangGraph application on LangGraph Cloud, your application's code must be stored in a GitHub repository. You can deploy any LangGraph applications to LangGraph Cloud with ease.

For this guide, we'll use the pre-built Python ReAct Agent template. You can go to GitHub and fork the repository. This ReAct Agent application requires API keys from Anthropic and Tavily.

Deployment to LangGraph Cloud

After logging in LangSmith, you can click the LangGraph Platform menu at the bottom of the left sidebar.
Click + New Deployment button at the bottom of the page and then you can follow the steps below for creating a new deployment.
- Select Github react-agent repository from the drop-down menu.
- Write the deployment name in Name field.
- Select Git branch. main is default.
- Langgraph config file is langgraph.json as default. You can also select another file.
- Select Development type.
- Write Environment Variables. In this tutorial, we will use ANTHROPIC_API_KEY and TAVILY_API_KEY.
- Click Submit button at the upper right corner. It takes a few minutes to build the application.
Now you can see the deployment status on the Overview section.

Using LangGraph Studio on the web

Once your application is deployed, you can test it in LangGraph Studio on the web.

You can find the LangGraph Studio text and Endpoint URL at the bottom of the page. Let's click the LangGraph Studio text to copy the clipboard.

Now you can test your LangGraph application in LangGraph Studio on the web.

Testing the API

Now we will send messages to the assistant using the Python SDK. you can also use JavaScript SDK or Rest API.

Prior to this, we need to install the langgraph-sdk package.

pip install langgraph-sdk or poetry add langgraph-sdk

from langgraph_sdk import get_client

# Initialize the LangGraph client with the endpoint URL and API key
client = get_client(url="{your_endpoint_url}", api_key="{your_langsmith_key}")

# Stream the response from the LangGraph assistant
async for chunk in client.runs.stream(
    None,       # Run without a specific thread (threadless run)
    "agent",    # Name of assistant. Defined in langgraph.json.
    input={
        "messages": [{
            "role": "human",    # User's role in the conversation
            "content": "say hello in french",   # User's input message
        }],
    },
    stream_mode="updates",  # Stream mode to receive updates in real-time
):
    print(chunk.data)   # Print the streamed data (assistant's response)

{'run_id': '1efef2b1-a250-6226-a457-8760b437980b', 'attempt': 1}
    {'call_model': {'messages': [{'content': 'To say "hello" in French, you don\'t need any special tools or searches. The French word for "hello" is:\n\n"Bonjour"\n\nThis is the most common and formal way to say hello in French. It literally translates to "good day" and can be used at any time of day.\n\nThere are also other ways to greet someone in French, depending on the time of day or the level of formality:\n\n1. "Salut" - A more casual way to say "hi" or "hey"\n2. "Bonsoir" - Used in the evening, meaning "good evening"\n3. "Coucou" - Very informal, similar to "hey there" or "hi there"\n\nIs there anything else you\'d like to know about French greetings or language?', 'additional_kwargs': {}, 'response_metadata': {'id': 'msg_01L2AGDYYLD8ADWyuAonxk55', 'model': 'claude-3-5-sonnet-20240620', 'stop_reason': 'end_turn', 'stop_sequence': None, 'usage': {'cache_creation_input_tokens': 0, 'cache_read_input_tokens': 0, 'input_tokens': 430, 'output_tokens': 181}}, 'type': 'ai', 'name': None, 'id': 'run-885a1808-0a9a-4763-a761-91972062d8c6-0', 'example': False, 'tool_calls': [], 'invalid_tool_calls': [], 'usage_metadata': {'input_tokens': 430, 'output_tokens': 181, 'total_tokens': 611, 'input_token_details': {'cache_read': 0, 'cache_creation': 0}}}]}}

Now, you can verify the message data.

If you append /docs to the end of the Endpoint URL and enter it in a web browser, you can check the web API. We can refer to this document and use API testing tools like Postman or Scalar to conduct tests.

ex) GET https://{{endpoint_url}}threads/{{thread_id}}/history

PreviousLangGraph Code Assistant NextTree of Thoughts (ToT)

Last updated 1 month ago