Deployment Options

Deploy OpenClaw using various methods

OpenClaw can run on your own hardware, in containers, on serverless runtimes, or as a managed service. The right choice depends on how much you want to operate yourself, where your data must live, and which features you need (e.g. browser automation and long-running cron jobs are not available in all environments).

When to run locally or in Docker: You need full control, browser automation, exec tools, or heavy skills. A home machine, VPS, or Docker on a server gives you the complete OpenClaw stack. Use the install guide for a single machine or the Docker guide for a reproducible, portable setup. For a guided flow that generates install or config scripts in your browser, see the Setup Wizard.

Local vs VPS: Running OpenClaw on a local device (same network as you) instead of a remote VPS often improves productivity: you can airdrop or paste files, use local paths, and get lower latency. Workflows like “drop a video, get transcript and thumbnails” are much easier when the Gateway and workspace are on a machine you use daily. Use a VPS when you need 24/7 reachability or don’t have suitable local hardware. More in Tips & best practices.

When to consider serverless (Moltworker): You want minimal ops and are okay with a subset of features—no browser, no host exec, no persistent local state. Moltworker runs the agent on Cloudflare Workers; good for simple chat and API-driven workflows.

When to use managed hosting (Clawi): You want WhatsApp, Telegram, or Discord connected without running any infrastructure. Clawi hosts OpenClaw for you; you configure channels and models in their dashboard. Best if your priority is getting an assistant live quickly without servers.

Want a hands-on tutorial? Follow our step-by-step Deployment Tutorial to learn how to deploy OpenClaw with Docker. This page provides an overview of deployment options.

🖥️ VPS Deployment

Run OpenClaw on a virtual server—Hetzner, DigitalOcean, Vultr, Hostinger. Full control, 24/7 uptime. Compare providers and follow step-by-step setup guides.

VPS Guide →

🐳 Docker Deployment

Deploy OpenClaw using Docker containers. Learn containerization, orchestration, and best practices for Docker-based deployments.

Docker Guide →

☁️ Serverless Deployment

Deploy OpenClaw on serverless platforms like Cloudflare Workers using Moltworker. Learn about serverless architecture and limitations.

Moltworker Guide →

🏠 Managed Cloud

Run OpenClaw without managing servers. Clawi, xCloud, MyClaw.ai, OpenClaw Cloud—WhatsApp, Telegram, Discord in minutes. No VPS, no SSH.

Managed Hosting Guide →

📦 BYOS (Bring Your Own Server)

You provide the VPS; a provider installs and manages OpenClaw. OpenClawHosting.io, MoltBotHost, Elest.io, ClawRun.dev. Data on your chosen infra.

BYOS Guide →

🏢 Enterprise & Teams

Multi-agent teams, shared workspaces, and team deployment. SSO, audit logs, and scaling patterns for organizations.

Enterprise →

Enterprise & Team Deployment

For teams and organizations, OpenClaw can run as multiple agents with shared context, per-channel sandboxing, and controlled DM access. Key patterns:

  • Multi-agent squad — Separate agents for different roles (support, research, ops). Multi-agent team tutorial
  • Shared context — One workspace or symlinked directory so agents read from the same page. Shared context use case
  • Sandboxing — Set agents.defaults.sandbox.mode to "non-main" so group sessions run in isolated containers
  • DM policy — Keep pairing for unknown DMs; use allowFrom to restrict who can message

For managed enterprise options, see Managed Hosting (some providers offer team plans). For self-hosted, use VPS or Docker with multiple Gateway instances or session routing.

When moving to a new server or reinstalling, back up config, credentials, and workspace—see Workspace Backup & Restore.