OpenClaw infrastructure
OpenClaw Hosting: Best VPS, Managed, Docker and Local Options
OpenClaw is a self-hosted AI assistant gateway that connects 32+ messaging platforms to AI coding agents. It runs as a single Node.js process on macOS, Linux, or Windows, with official Docker, Podman, and Kubernetes support. There is no single best hosting model for every deployment: a VPS favors control and predictable 24/7 availability, managed or container platforms reduce some server operations, and local hardware favors direct control and privacy. Provider-level rankings remain test pending.
Which OpenClaw hosting option is best?
Choose by operating model, not by roundup rank. Use a Linux VPS when you want an always-on gateway with full server control; use a managed or container platform when you want to reduce host administration; use local hardware when privacy, LAN access, or colocated local inference matters. AIG's provider tests will add measured uptime, restarts, RAM use, and total monthly cost before any specific hosting provider is labeled a winner.
What is OpenClaw?
OpenClaw is an open-source (MIT) AI assistant gateway developed by the OpenClaw Foundation, an independent 501(c)(3) nonprofit. It has no paid tier and no hosted service — you run it on your own infrastructure. The gateway connects messaging platforms (Discord, Slack, Telegram, WhatsApp, iMessage, Signal, Matrix, and others) to AI models from OpenAI, Anthropic, Ollama, LM Studio, OpenRouter, and other providers.
| Field | Details |
|---|---|
| Runtime | TypeScript on Node.js 24.16+ or 26.1+ (26 recommended) |
| License | MIT — OpenClaw Foundation (501(c)(3)) |
| Platforms | macOS, Linux, Windows, ChromeOS, iOS, Android |
| Default gateway port | 18789 (loopback only) |
| Configuration | ~/.openclaw/ (overridable via OPENCLAW_HOME) |
| Container support | Docker, Podman, Kubernetes |
| Model providers | OpenAI, Anthropic, Ollama (auto-discovery), LM Studio, OpenRouter, DeepInfra, Cerebras, custom endpoints |
| Messaging | 32+ platforms via channel plugins |
| Current version | v2026.9.4 (September 2026) |
Preparing a persistent gateway? Use the free OpenClaw security and host-readiness checker to review private gateway access, restart recovery, backups, monitoring, and cost controls. This checklist is self-reported; it does not perform remote host testing.
OpenClaw System Requirements
OpenClaw requires Node.js 24.16+ or 26.1+ (the installer provisions Node 26 automatically when needed). The official documentation does not publish minimum RAM or CPU requirements. Agent Infra Guide will add measured resource baselines after completing hands-on deployment tests.
OpenClaw Installation Options
| Method | Platform | Notes |
|---|---|---|
| Shell installer | macOS, Linux, WSL2 | Recommended. Detects OS, installs Node if needed. |
| PowerShell installer | Windows | Native Windows installation path. |
| npm / pnpm / bun | Any | Package manager install when Node is already present. |
| Docker | Any Docker host | Official images. Container defaults to exposed bind with auth. |
| Podman | Linux | Rootless container alternative. |
| Kubernetes | Cluster | Deployment configurations available. |
| macOS menu-bar app | macOS | Desktop companion with LaunchAgent service. |
| Windows Hub | Windows | Signed desktop installer. |
OpenClaw Hosting and Deployment Options
The right hosting option depends on whether you need the gateway available 24/7, whether you want to run local models alongside the agent, and how much infrastructure you want to manage.
| Option | Good fit | Tradeoff | OpenClaw docs |
|---|---|---|---|
| Linux VPS | Persistent 24/7 gateway with cloud API models | You manage the host OS, security, and updates | Guides for Hetzner, DigitalOcean, GCP, Azure |
| Docker on VPS | Repeatable containerized deployment | Container and host security both matter | Official Dockerfile + docker-compose.yml |
| Container platform | Lower operational burden | Less control over persistence and networking | Guides for Fly.io, Railway |
| Local Mac / Linux | Privacy, direct access, local models via Ollama | You own uptime and connectivity | LaunchAgent (macOS), systemd (Linux) |
| Kubernetes | Existing cluster, multi-service environment | Operational complexity beyond what most agents need | Deployment configs available |
OpenClaw Network and Gateway Security
The OpenClaw gateway binds to loopback (127.0.0.1) by default on normal host installations. The dashboard is available at http://127.0.0.1:18789/. Container images default to an exposed bind but require authentication.
Before exposing the gateway beyond loopback, the documentation requires completing a pre-flight and rollout checklist. Supported remote-access methods include SSH tunneling, Tailscale, and reverse proxy with trusted proxy authentication. The gateway assumes one trust boundary per instance — it is designed for a single operator or trusting team, not adversarial multi-tenant use.
Run openclaw security audit to detect drift from secure defaults after installation or configuration changes.
OpenClaw Local Model Support
OpenClaw includes Ollama as an official provider. When Ollama runs on the same machine as the gateway, it is discovered automatically — no manual endpoint configuration is needed. The official documentation includes dedicated pages for Ollama setup, configuration recipes, vision support, and node-local inference.
LM Studio and other OpenAI-compatible local model servers are also supported via custom endpoint configuration.
VPS Plans — Candidates for AIG Testing
Agent Infra Guide has not yet tested OpenClaw on specific VPS plans. The providers below have documented pricing and are candidates for hands-on deployment tests. None should be treated as a tested recommendation until AIG publishes measured results.
| Provider | Plan | RAM | Price | Notes |
|---|---|---|---|---|
| DigitalOcean | Basic Droplet | 2 GB | $12/mo | 1 vCPU, 50 GB SSD. OpenClaw has an official deployment guide. |
| Vultr | Cloud Compute | 2 GB | $10/mo | 1 vCPU, 55 GB SSD. Hourly billing. |
| Hetzner | Cloud (ARM or x86) | 4 GB | Verify current price | 2 vCPU, 40 GB NVMe. OpenClaw has an official Docker deployment guide. |
| Fly.io | Shared CPU | 2 GB | ~$11/mo | Per-second billing. OpenClaw has an official deployment guide. Storage extra. |
| Hostinger | KVM 1 | 4 GB | $6.49/mo (promo) | Promotional price; renews at $11.99/mo. 2-year prepaid. |
Sources
- OpenClaw — Installation— Official installation methods, platform support, and runtime requirements
- OpenClaw — Gateway Security— Default network binding, authentication, remote access, and security audit
- OpenClaw — Ollama Configuration— Official Ollama provider setup, auto-discovery, and local model configuration
- OpenClaw GitHub Repository— Source code, Docker configuration, and release history
Agent Infrastructure Recommendation Kit
Turn OpenClaw hosting research into a complete infrastructure plan
The Recommendation Kit helps you document workload requirements, compare VPS, local, Docker, and hybrid deployment options, record current provider pricing and evidence, and validate the architecture before deploying OpenClaw to production.
Get the Recommendation Kit — $19