CLI Commands

Core command-line interface providing all LangGraph development and deployment workflows. All commands support --help for detailed option information.

Capabilities

Project Creation

Create new LangGraph projects from predefined templates.

langgraph new [PATH] --template TEMPLATE_NAME

Parameters:

• PATH (optional): Directory path for new project
• --template: Template identifier (see Templates documentation)

Usage Examples:

# Create project in current directory with interactive template selection
langgraph new

# Create project in specific directory
langgraph new ./my-langgraph-app

# Create project with specific template
langgraph new my-project --template react-agent-python

Development Server

Run LangGraph API server in development mode with hot reloading and debugging capabilities.

langgraph dev [OPTIONS]

Key Options:

• --host TEXT: Network interface to bind (default: 127.0.0.1)
• --port INTEGER: Port number (default: 2024)
• --no-reload: Disable automatic reloading
• --config, -c FILE: Configuration file path (default: langgraph.json)
• --n-jobs-per-worker INTEGER: Max concurrent jobs per worker
• --no-browser: Skip opening browser window
• --debug-port INTEGER: Enable remote debugging on port
• --wait-for-client: Wait for debugger connection
• --studio-url TEXT: LangGraph Studio URL
• --allow-blocking: Don't raise errors for blocking I/O
• --tunnel: Expose via public tunnel
• --server-log-level TEXT: API server log level (default: WARNING)

Usage Examples:

# Basic development server
langgraph dev

# Custom port and host
langgraph dev --host 0.0.0.0 --port 3000

# Development with debugging
langgraph dev --debug-port 5678 --wait-for-client

# Disable auto-reload and browser opening
langgraph dev --no-reload --no-browser

Production Deployment

Launch LangGraph API server using Docker Compose for production deployment.

langgraph up [OPTIONS]

Key Options:

• --port, -p INTEGER: Port to expose (default: 8123)
• --config, -c FILE: Configuration file path (default: langgraph.json)
• --docker-compose, -d FILE: Additional docker-compose.yml file
• --recreate/--no-recreate: Container recreation control (default: no-recreate)
• --pull/--no-pull: Pull latest images (default: pull)
• --watch: Restart on file changes
• --wait: Wait for services to start
• --verbose: Show detailed server logs
• --debugger-port INTEGER: Serve debugger UI on port
• --debugger-base-url TEXT: URL for debugger to access API
• --postgres-uri TEXT: Custom PostgreSQL connection string
• --api-version TEXT: API server version to use
• --image TEXT: Pre-built Docker image to use
• --base-image TEXT: Base image for the server

Usage Examples:

# Basic production deployment
langgraph up

# Custom port and configuration
langgraph up --port 9000 --config my-config.json

# With additional services and debugging
langgraph up --docker-compose docker-compose.services.yml --debugger-port 8080

# Force recreate containers with verbose logging
langgraph up --recreate --verbose --wait

Image Building

Build Docker images for LangGraph applications for deployment or distribution.

langgraph build -t IMAGE_TAG [OPTIONS] [DOCKER_BUILD_ARGS]...

Key Options:

• --tag, -t TEXT: Docker image tag (required)
• --config, -c FILE: Configuration file path (default: langgraph.json)
• --pull/--no-pull: Pull latest base images (default: pull)
• --base-image TEXT: Override base image
• --api-version TEXT: API server version
• --install-command TEXT: Custom install command
• --build-command TEXT: Custom build command

Parameters:

• DOCKER_BUILD_ARGS: Additional Docker build arguments (passed through)

Usage Examples:

# Basic image build
langgraph build -t my-app:latest

# Multi-platform build with custom base image
langgraph build -t my-app:v1.0 --base-image python:3.11-slim --platform linux/amd64,linux/arm64

# Build with custom configuration and no base image pull
langgraph build -t my-app:dev --config dev.json --no-pull

# Advanced build with Docker build args
langgraph build -t my-app:latest --build-arg HTTP_PROXY=http://proxy.example.com:8080

File Generation

Generate Dockerfile and Docker Compose files for custom deployment workflows.

langgraph dockerfile SAVE_PATH [OPTIONS]

Parameters:

• SAVE_PATH: Path where Dockerfile will be saved (required)

Key Options:

• --config, -c FILE: Configuration file path (default: langgraph.json)
• --add-docker-compose: Generate docker-compose.yml and related files
• --base-image TEXT: Override base image
• --api-version TEXT: API server version

Usage Examples:

# Generate Dockerfile only
langgraph dockerfile ./Dockerfile

# Generate Dockerfile and Docker Compose files
langgraph dockerfile ./Dockerfile --add-docker-compose

# Generate with custom configuration and base image
langgraph dockerfile ./deploy/Dockerfile --config prod.json --base-image python:3.12-alpine

Global Options

All commands support these global options:

• --help: Show help message and available options
• --version: Show CLI version information

Configuration File

All commands default to using langgraph.json in the current directory. This file defines:

• dependencies: Python/Node.js packages required
• graphs: Mapping of graph IDs to import paths
• env: Environment variables (file path or dictionary)
• python_version: Python runtime version (3.11, 3.12, 3.13)

Example minimal configuration:

{
  "dependencies": ["langchain_openai", "."],
  "graphs": {
    "my_graph": "./src/graph.py:compiled_graph"
  },
  "env": "./.env"
}

Error Handling

The CLI provides detailed error messages for common issues:

• Configuration validation errors: Invalid langgraph.json format or missing required fields
• Docker errors: Missing Docker/Docker Compose or insufficient permissions
• Port conflicts: Port already in use for development or production servers
• Template errors: Invalid or missing template specifications
• Build failures: Docker build errors with context and suggestions

Use --verbose flag with applicable commands for additional debugging information.