create-agent-chat-app

A CLI tool to quickly set up a LangGraph agent chat application.

This will clone a frontend chat application (Next.js or Vite), along with up to 4 pre-built agents. You can use this code to get started with a LangGraph application, or to test out the pre-built agents!

Usage

Clone code:

# Using npx
npx create-agent-chat-app@latest
# or
yarn create agent-chat-app@latest
# or
pnpm create agent-chat-app@latest
# or
bunx create-agent-chat-app@latest

You'll then be prompted for the name of the project, the package manager, the web framework, and which, if any, agents to include by default:

✔ What is the name of your project?
✔ Which package manager would you like to use? › npm | pnpm | yarn
✔ Would you like to automatically install dependencies? … y / N
✔ Which framework would you like to use? › Next.js | Vite
✔ Would you like to include all pre-built agents? … y / N

If you select y for the question Would you like to include all pre-built agents?, it will include the following agents:

• React Agent
• Memory Agent
• Research Agent
• Retrieval Agent

If you select N, it will prompt you one by one to select which agents to include.

✔ Include ReAct agent? › y / N
✔ Include Memory agent? › y / N
✔ Include Research agent? › y / N
✔ Include Retrieval agent? › y / N

After you finish the prompts, it will automatically create all the necessary files and folders in the project directory. If you selected auto-install dependencies, it will install them for you.

Setup

Navigate into the project directory:

# agent-chat-app is the default project name
cd agent-chat-app

Copy the .env.example file to .env:

cp .env.example .env

This will contain all of the required secrets the agent(s) need in order to run.

Finally, start the development servers. This command will start both the web, and LangGraph servers:

npm run dev
# or
pnpm dev
# or
yarn dev

If you choose to run them independently, you can by either running the Turbo command from the root of the project:

Web:

npm turbo dev --filter=web
# or
pnpm turbo dev --filter=web
# or
yarn turbo dev --filter=web

LangGraph:

npm turbo dev --filter=agents
# or
pnpm turbo dev --filter=agents
# or
yarn turbo dev --filter=agents

Or, you can navigate into each workspace, and run dev:

Web:

cd apps/web

npm run dev
# or
pnpm dev
# or
yarn dev

LangGraph:

cd apps/agents

npm run dev
# or
pnpm dev
# or
yarn dev

Once the server is running, you can visit http://localhost:3000 (or http://localhost:5173 for Vite) in your browser. From there, you'll be prompted to enter:

• Deployment URL: The API URL of your LangGraph server. You should use the default value of http://localhost:2024, as this is what the LangGraph server which ships with this package is configured to run on.
• Assistant/Graph ID: The name of the graph, or ID of the assistant to use when fetching, and submitting runs via the chat interface. If you selected the ReAct agent, you can use the default value of agent to connect to it. Otherwise, consult the langgraph.json file to find the graph ID of the agent you would like to connect to.
• LangSmith API Key: This field is not required for local development. Your LangSmith API key to use when authenticating requests sent to LangGraph servers.

After entering these values, click Continue. You'll then be redirected to a chat interface where you can start chatting with your LangGraph server.

Why use Create Agent Chat App?

This tool is a quick way to get started with a LangGraph chat application. It is based off of the Agent Chat UI repository, and ships by default with 4 pre-built agents.

Using the Agent Chat UI, you're able to interact, and chat with these agents.