Hostinger OpenClaw Setup

Hostinger offers three ways to set up OpenClaw, and picking the right one before you start saves youslug from redoing checkout.

This guide compares all three, then walks through the one-click VPS method step by step, including the setup choice that decides whether your assistant is ready to chat immediately or needs one extra step first, plus the security settings, cost tradeoffs, and troubleshooting details most walkthroughs skip.

Choose Your Setup Method

Pick Managed OpenClaw if you want to skip Docker, environment variables, and server administration entirely. Hostinger handles the container, updates, and infrastructure. You connect an AI provider and a messaging channel, and that’s the extent of the setup. One subscription also includes Hermes Agent, a separate Hostinger AI tool, at no extra cost.

Pick OpenClaw on VPS if you want root access and direct control over the Docker container without building it from scratch yourself. Hostinger deploys the container automatically the moment your server provisions, but everything past that, environment variables, security settings, additional containers, is yours to configure. This matters in practice if you ever want to run a second service alongside OpenClaw on the same server, inspect logs directly, or move to a different AI provider without waiting on a managed platform to support it. This is the method covered in the walkthrough below.

Pick a manual VPS install if you want to build OpenClaw from the ground up: provision a bare server, install Docker yourself, clone the repository, and run the setup wizard from the terminal. This gives you the same end result as the one-click template, with every automated step done by hand instead. Choose this only if you’re deploying outside Hostinger’s one-click options or want full visibility into the build process from the first command.

Deploy OpenClaw on a Hostinger VPS

Launch OpenClaw with Hostinger's one-click VPS setup and get full root access, Docker control, and the flexibility to connect your preferred AI provider.

Visit Hostinger

How to Set Up OpenClaw on a Hostinger VPS

I’m walking through this method because it sits at the most useful middle ground of the three. Managed OpenClaw hides the container entirely, so there’s nothing to actually show you past a few clicks.

A fully manual install means cloning a repository and running setup commands in a terminal before you see anything at all.

The VPS template gives you the real environment, root access, Docker containers, security settings you can actually open and change, without making you build that environment from scratch first. If you want to understand what’s running under an AI assistant rather than just trusting a managed dashboard, this is the version to see step by step.

Step 1: Choose Your VPS Plan

Go to Hostinger’s OpenClaw page and click Choose plan.

Four KVM tiers are available:

Sizing guidance: OpenClaw itself is a lightweight Node.js process, so the plan you need depends less on the assistant and more on what you connect to it:

• KVM 1 covers a single user messaging through one or two channels with occasional automations. This is enough for personal use.
• KVM 2 (the plan marked Most Popular) gives real headroom if you’re running several automations, connecting multiple channels at once, or planning to add a second Docker container alongside OpenClaw later.
• KVM 4 or higher only makes sense if you’re running OpenClaw for a team, handling high message volume across several channels simultaneously, or pairing it with resource-heavy skills like browser automation running frequently.

Step 2: Set Up Your Cart

Three add-ons appear under “OpenClaw auto-deploys with your VPS”:

• Ready to use AI: pre-purchased nexos.ai credits connecting OpenClaw to Anthropic, OpenAI, Google, and xAI models with no API key setup required

• Instant web scraping: free Oxylabs credits for real-time web access

• Agentic Mail: a dedicated email address for the assistant, billed separately

The AI credits decision is the one that matters most. Here’s the actual tradeoff:

If you bring your own key, note that both Anthropic and OpenAI require an active billing method on file before the API will respond to requests.

A key without billing configured will fail the same way a missing key does, so if Step 7 doesn’t work after you’ve added the variable, check your provider’s billing dashboard before assuming Hostinger’s side is broken.

• If you check “Ready to use AI” and pay for the credits, OpenClaw is ready to chat the moment setup finishes.
• If you leave it unchecked to use your own API key instead, you’ll need to add that key manually after deployment, covered in Step 7 below.

Pick a server location from the dropdown, which shows a latency estimate for each region. Location affects response speed for the web dashboard and any channel connections routed through it, so pick the region closest to where you’ll actually be using the assistant, not necessarily where the server itself is cheapest.

Step 3: Register and Pay

Register with Google, GitHub, or email and password. On the payment screen, choose from card, PayPal, Google Pay, AliPay+, or Coingate. Payment processes on the same screen, no separate redirect.

Step 4: Open Your Server in hPanel

After payment, you’ll land in hPanel automatically. Scroll to the VPS section to find your new server, marked Running with its IP address visible. Click Manage to open its dashboard.

Step 5: Copy Your Gateway Token

The VPS Overview page opens on a My applications card showing OpenClaw already deployed and running as a single container. Underneath the app name:

• App terminal: direct terminal access to the container
• Documentation: links to OpenClaw’s official docs
• Gateway token: click the copy icon to copy it

This token logs you into OpenClaw’s own web interface. It’s separate from your Hostinger account password, and it functions as a full authentication key to your assistant, not a throwaway code.

Click Open app on the same card to go to the login screen.

Step 6: Log In

Paste your gateway token into the login field and click Login.

What happens next depends on your Step 2 choice:

• Paid for AI credits: You land in a working chat interface.
• Skipped the credits: You’ll see “AI Provider API keys are missing.” This isn’t an error, it’s expected. Continue to Step 7.

Get Full Control of OpenClaw With Hostinger

Hostinger's OpenClaw VPS template gives you a pre-deployed Docker environment while keeping root access, environment variables, and server configuration under your control.

Visit Hostinger

Step 7: Add Your Own API Key

Skip this step if you already have a working chat interface.

1. In hPanel, open your VPS and click Docker Manager in the sidebar, then Applications.

2. You’ll see two running containers: OpenClaw and traefik. Traefik is a reverse proxy, it’s what routes traffic from the internet to the right container and handles the TLS certificate for your connection. You won’t need to touch it directly, but it helps to know it’s there if you ever see connection errors that aren’t coming from OpenClaw itself.
3. Click Manage next to the OpenClaw container. This opens a visual environment variable editor, not a raw config file.
4. Existing variables include PORT, TZ, OPENCLAW_GATEWAY_TOKEN, and TRAEFIK_HOST, all set automatically during deployment.
5. Click + Environment and add one of the following, depending on your provider:

6. Paste your actual API key as the value.
7. Click Save and deploy. The container rebuilds and restarts automatically, no manual restart needed.

Step 8: Log Back In and Confirm Your Model Is Active

Once Docker Manager finishes redeploying the container, go back to the OpenClaw web interface before doing anything else.

Restarting a container clears whatever was running inside it, including the session you were logged into, so the login screen you saw in Step 6 will be waiting for you again rather than the dashboard you left.

Paste your gateway token into the field and click Login, the same token you copied back in Step 5. If you saved it somewhere retrievable, this takes a few seconds. If you didn’t, head back to the VPS Overview page in hPanel and copy it again from the same My applications card.

This time, instead of the “AI Provider API keys are missing” screen from Step 6, you should land directly in a working dashboard: an assistant avatar, a status badge reading Ready to chat, four suggested prompts sitting above the message box, and a settings gear in the bottom corner.

That dashboard loading correctly is your confirmation the key you added in Step 7 actually took effect. If you still see the missing-key warning after logging back in, double-check that the variable name in Docker Manager exactly matches one from the table in Step 7, a typo in the name field is the most common reason the key doesn’t get picked up.

Before you send a real message, check one more thing: the model selector in the bottom right of the chat screen.

Adding a key doesn’t always switch OpenClaw over to using it automatically, so if that selector still shows a different provider than the one you just connected, open Settings and change the active model manually under Model & Thinking.

This screen also lets you set a Thinking level (Off, Low, Medium, High) and a Fast mode, both of which trade response speed against reasoning depth. Adjust these if you notice replies feeling either too slow or too shallow for what you’re asking.

Step 9: Connect a Messaging Channel

None of the six supported channels connect automatically, regardless of which setup method you use. What each one involves varies:

• Telegram is usually the fastest to connect: it uses a bot token generated through Telegram’s own BotFather, then paired to your account with a short code.
• WhatsApp pairs through a QR code shown in the OpenClaw interface, similar to linking WhatsApp Web.
• Discord and Slack connect through each platform’s own bot or app authorization flow, which grants OpenClaw permission to read and respond in the channels you choose.
• Signal and iMessage have narrower support and are usually the most involved to set up, since neither platform offers the same kind of open bot API the others do.

Go to Settings > Channels and click Connect next to whichever platform you want.

Send a test message once it’s connected to confirm the assistant, the channel, and your AI provider are all actually working together before you rely on it.

Step 10: Review Security and Resource Settings

Before you consider the setup finished, check these sections in Settings.

Security.

The Tool profile setting deserves real attention. “Full” grants access to every available tool, including anything installed later.

If you’re running OpenClaw for one specific job, narrowing this to “minimal” or “messaging” limits what the assistant can actually do if a conversation ever goes somewhere you didn’t intend, the same principle as giving an application the narrowest permissions it needs rather than the broadest available.

Gateway Host. Shows live server specs: OS and kernel version, Node.js runtime version, uptime, CPU load, memory used, and disk space free. Check this periodically if OpenClaw starts responding slowly, since a server running close to its memory or CPU limit is a common, quiet cause of a sluggish assistant.

Context Profile. Four presets control how much workspace context loads per interaction, which directly affects both cost and response quality if you’re paying per token through your own API key:

A larger context budget usually means better answers on complex requests, at the cost of more tokens billed per interaction if you’re on your own API key rather than flat-rate credits.

Run Your AI Assistant on Hostinger VPS

Use Hostinger VPS to run OpenClaw with root access, Docker Manager, configurable AI providers, messaging channels, and security controls in one environment.

Visit Hostinger

Keeping OpenClaw Running Reliably

A few habits matter more once OpenClaw is live than they did during setup:

• Back up your gateway token and API keys somewhere separate from the server. If the VPS ever needs to be rebuilt, you’ll need both to reconnect without starting over.
• Watch resource usage if the assistant handles frequent messages or scheduled automations. The CPU, memory, and disk figures on the Gateway Host screen are the fastest way to catch a server that’s about to run out of headroom.
• Update periodically. Docker Manager will show a newer image is available when one exists. Treat OpenClaw updates the same as any other server software: apply them for security fixes, but check release notes first if you have custom automations that might depend on current behavior.
• Don’t skip billing setup with your AI provider if you’re using your own key. A key with no payment method attached fails silently in a way that looks identical to a misconfigured environment variable.

Quick Reference: What to Do If…

Which Method Should You Pick?

If you’ve never touched Docker, choose Managed OpenClaw. If you want root access and don’t mind an extra step when you skip the paid AI credits, use the one-click VPS template covered above. Save a fully manual install for when you’re deploying outside Hostinger entirely or want to understand OpenClaw’s build process from the first command.

Hostinger

5.0

2,471 user reviews

£1.17 /mo

Starting price

Visit Hostinger

Rating based on expert review

• User Friendly
  4.9
• Support
  4.9
• Features
  4.8
• Reliability
  4.7
• Pricing
  4.5