Install OpenClaw without the dependency headaches

OpenClaw requires a compatible runtime environment, sufficient system resources, and network connectivity to model provider endpoints. Understanding these requirements before starting prevents mid-installation failures that require cleanup and restart.

For macOS installations, the system must run macOS 12 Monterey or later with Apple Silicon (M1, M2, M3, or M4 chips). Intel-based Macs are supported but perform best with macOS 13 or later. The installer requires approximately 2GB of available disk space for the runtime environment and working files.

Windows installations require Windows 11 with the latest cumulative updates installed. Windows 10 is supported for development environments but is not recommended for production deployments due to dependency resolution limitations in the Windows package management ecosystem.

Docker installations require Docker Desktop 4.0 or later with at least 4GB of memory allocated to containers. The Docker path is recommended for teams that need reproducible environments across multiple developer machines or CI/CD pipelines.

The macOS installation path for OpenClaw targets Apple Silicon hardware with runtime optimizations specific to the ARM architecture. The guided installer handles Homebrew installation if missing, verifies node runtime version compatibility, and configures local data directories with appropriate permission scopes.

Start by downloading the latest macOS installer package from the official release page. Run the installer with administrator privileges when prompted for permission to write to system directories. The installer will automatically detect whether Homebrew is present and install it with required dependencies if absent.

After installation completes, verify the setup by running the health check command in Terminal. The health check validates node runtime connectivity, confirms data directory permissions, and tests network access to model provider endpoints. Successful health check output indicates the agent is ready for configuration.

Common installation failures on macOS include Homebrew path conflicts when multiple package managers are installed, permission errors when the installer runs from a network-mounted volume, and version mismatches when upgrading from an existing OpenClaw installation. Each failure mode has a specific recovery procedure documented in the troubleshooting section.

The Windows installation path for OpenClaw provides native binaries that bundle runtime dependencies, eliminating the need for WSL or separate node installation. This approach reduces installation failures caused by environment configuration drift across Windows versions and configurations.

Download the Windows installer executable and run it with local administrator privileges. The installer will prompt for installation directory selection and offer options for adding OpenClaw to the system PATH for command-line access. Accept the defaults unless your organization requires non-standard installation paths.

Post-installation verification requires opening a new PowerShell or Command Prompt window and running the health check command. The verification process tests runtime connectivity, confirms file system permissions for the installation directory, and validates network access to external model provider endpoints.

Windows-specific installation issues typically involve antivirus interference with the installer binary, path length limitations in deeply nested directory structures, and permission errors when User Account Control settings are configured for maximum security. The troubleshooting guides address each of these scenarios with specific remediation steps.

Docker Compose deployment packages OpenClaw with its complete runtime environment in a containerized configuration that teams can share and reproduce consistently. This approach is ideal for development teams, CI/CD pipelines, and scenarios where multiple developers need identical environments without individual installation complexity.

Clone the official Docker Compose template repository or download the compose file directly. The template includes service definitions for the agent core, optional PostgreSQL database for state persistence, and volume mounts for configuration and log files that persist across container restarts.

Before starting the containers, verify Docker Desktop is running with sufficient memory allocation. The minimum configuration requires 4GB of memory and 2 CPU cores for acceptable performance during task execution. Edit the docker-compose.yml file to adjust resource allocation for your specific hardware capabilities.

Start the containers using docker compose up -d in the directory containing the compose file. Monitor container health using docker compose ps to confirm all services reach healthy status. Access the agent management interface through the exposed port on localhost to complete initial configuration.

Local installations suit individuals and small teams evaluating the platform, workflows that require local data processing without network transmission, and scenarios where persistent cloud infrastructure costs are difficult to justify for intermittent usage patterns.

Hosted installation through VPS deployment eliminates local hardware requirements, provides consistent uptime without local machine maintenance, and offers centralized logging that simplifies troubleshooting when issues arise. The hosted path is recommended for production workflows and team-based deployments where multiple users need concurrent access.

Both paths use identical agent core software, so configuration and operational procedures transfer directly between deployment modes. Teams commonly start with local installation for evaluation, then migrate to hosted deployment as usage scales and reliability requirements increase.

Completing the installation without health verification is the most common cause of confusing failures during subsequent configuration steps. The health check command confirms all runtime dependencies are present, network connectivity is functional, and the agent can reach model provider endpoints before you invest time in configuration.

Run the health verification command immediately after installation, before proceeding to model configuration or skills activation. Successful verification produces output indicating each subsystem is operational. Failures at this stage indicate environmental issues that will block all agent functionality.

After health verification passes, run a simple validation task to confirm the agent can execute workflows end-to-end. The validation task should be something simple like fetching current weather information or retrieving a public webpage title. Successful validation confirms the installation is complete and ready for production configuration.