Before you start
Before you deploy your first AI agent, make sure you have the following ready. The build itself takes about 25 minutes for OpenClaw and five to seven minutes for Hermes Agent, and you can close the tab while it runs.
- A OneClickClaw account. If you have not signed in yet, see Creating Your Account. It takes 30 seconds with Google Sign-In.
- An agent and a plan in mind. Pick OpenClaw or Hermes Agent. Most people start with the 7-day free trial, which needs no card and runs on a real Starter server. If you prefer to pay from day one, you pick a plan and pay through Stripe. To compare plans, see Choosing a Plan.
- An AI account. Your agent runs on an AI account that belongs to you. You can sign in with a ChatGPT plan (Plus or higher) or a SuperGrok or X Premium+ plan. Hermes Agent also accepts a Nous Portal or MiniMax plan. Or use an API key from a supported provider such as Anthropic, OpenAI, OpenRouter or xAI. Claude works with an Anthropic API key. For help, see BYOK Explained.
- A messaging channel (optional). If you want your agent to answer in Telegram or Discord, have your bot token ready. WhatsApp only needs your phone to scan a QR code. You connect channels inside your agent, now or later.
Tip
A common setup is: the free trial on Starter, the ChatGPT plan you already pay for, and a Telegram bot token. If you keep the server afterwards, Starter costs €14.99/month and your AI stays on the plan you already have.
The deployment flow
There are two ways in: the 7-day free trial, or a paid plan from day one. Both end on the same build page. Here is exactly what you see:
Start the free trial
On your empty dashboard, press Start free OpenClaw trial or Start free Hermes trial. The next page says "Start your 7-day free trial" for the agent you picked. Type a name for your server in Name Your Server (up to 60 characters, for example "Support Bot"). The name only appears in your dashboard, and you can rename the server later. Then press the start button below it. No card is needed. The build usually starts at once.
Name Your Server
Text field, up to 60 characters
"You can rename it later from the dashboard."
Start button
[Start free OpenClaw Starter trial]
Or pick a paid plan: choose your plan
If you want to pay from day one, press Or pick a paid plan on your dashboard. The first step, Choose Your Plan, shows Starter, Hobbyist and Pro with their specs and prices. Choose monthly or annual billing, pick a plan and press Continue. This button orders an OpenClaw server. A paid Hermes Agent server starts from the plan buttons on the Hermes Agent page.
Review, name your server and pay
The second step, Review & Deploy, shows your order. Type your server name in Name Your Server and add a coupon code if you have one. Press Proceed to Payment to open the secure Stripe payment page, enter your card and confirm. Stripe handles all payment processing. We never see or store your card details. Once payment is confirmed, the build starts automatically.
Name Your Server
Text field, up to 60 characters
Order summary
Plan: Starter, monthly or annual
Coupon code field
Button
[Proceed to Payment]
Watch the build (about 25 minutes for OpenClaw, five to seven for Hermes Agent)
Your dedicated server is created, configured, and secured automatically. The build page shows each step live. OpenClaw takes about 25 minutes and Hermes Agent five to seven minutes. You can stay on the page and watch, or close the tab: we email you the moment the server is ready, and the email's button opens your agent. On a Hermes Agent build, the page also asks you to choose the AI behind your agent while you wait.
Open your agent and connect your AI
When the build is done, press Open Your Agent (OpenClaw) or Open my agent (Hermes Agent). On OpenClaw, the agent opens on its Model Setup page: press Connect provider, then choose how to connect. For a ChatGPT plan, choose OpenAI, then Codex login (device code). For SuperGrok or X Premium+, choose xAI (Grok), then xAI OAuth. Or paste an API key. On Hermes Agent, if you chose a subscription on the build page and kept that page open, its sign-in opens by itself when the server is ready and shows a code: approve it with your AI account. You can connect or change the AI any time from the AI Model card on your server page. Then connect your chat channels inside the agent.
Note
A free trial runs for 7 days. Pay for Starter (monthly or annual) before the 7 days end to keep the same server and everything on it; after that you can move up to a larger plan. When the 7 days end without a payment, the server is paused and kept with its data for 3 more days; paying in that time switches the same server back on. After that it is removed with its data.
What happens during provisioning
Once you start the trial or your payment is confirmed, OneClickClaw creates and configures your dedicated server automatically. You do not need to do anything during this process. Here is what happens behind the scenes:
- Server creation. A new VPS is created in our EU datacenter (Denmark) with the CPU, RAM, and SSD specs matching your plan.
- Software installation. Ubuntu is prepared and hardened (firewall, SSH keys, automatic security updates). Then your agent is installed: OpenClaw with Node.js and its chat add-ons, or Hermes Agent with its own Python setup.
- Web address and final checks. SSL certificates are provisioned so all traffic is encrypted. Health monitoring is enabled. The server reports back that it is ready, and we email you.
Note
The whole build takes about 25 minutes for OpenClaw and five to seven minutes for Hermes Agent. Some servers take a little longer: a build that is still working is given extra time, so a slow build is no reason to worry. For more detail on what happens at each stage, see What Happens During Provisioning.
Creating your server
Checkmark when done
Preparing the system
Checkmark when done
Downloading OpenClaw
Checkmark when done
First-time setup and chat add-ons
Checkmark when done
Checking the settings
Spinner while it runs
Starting your agent
Waiting
Securing your web address
Waiting
Final checks
Waiting
You can leave this page
"We email you the moment it is ready."
Setup inside OpenClaw
When you open your agent for the first time, OpenClaw's own Model Setup page asks for your AI model. Press Connect provider and choose how to connect: for a ChatGPT plan (Plus or higher), choose OpenAI, then Codex login (device code); for SuperGrok or X Premium+, choose xAI (Grok), then xAI OAuth; or pick a provider and paste its API key. OpenClaw sends one real request to the provider and turns conversations on only after a model answered. A provider added from the Models page needs one more click: choose it in the Default model dropdown and save.
Chat channels are added under Settings, then Channels, each with its own guided setup. For a detailed walkthrough, see Set Up Your Agent Inside OpenClaw.
Note
On Hermes Agent, the AI is connected from OneClickClaw: the AI choice card on the build page, or the AI Model card on your server page. It takes a ChatGPT plan, SuperGrok or X Premium+, Nous Portal, a MiniMax plan, or an API key. After a sign-in, the server asks the model one short question before it tells you the AI is connected. Chat channels are connected inside the Hermes dashboard.
Tip
Web chat works with no channel at all: you can talk to your agent in its own web interface. Add WhatsApp, Telegram or Discord whenever you want the agent on your phone.
What to do after deployment
Congratulations, your AI agent is live! Here are the most valuable things to explore next:
- Customize your agent's personality. On OpenClaw, the Workspace Files card on your server page holds SOUL.md, where you define how your agent behaves, what tone it uses, what topics it focuses on, and what it refuses to discuss. A well-written personality is the difference between a generic chatbot and a useful assistant.
- Add more channels. Connect WhatsApp, Telegram, Discord, Slack and more inside your agent (on OpenClaw: Settings, then Channels). Your agent can serve several channels at the same time with the same personality and configuration.
- Monitor performance. Your dashboard shows real-time metrics including CPU usage, memory consumption, and uptime. Keep an eye on these during the first few days to make sure your plan is a good fit for your workload.
- Experiment with models. Try different AI models to find the best balance of quality, speed, and cost for your use case. On OpenClaw you switch under Settings, then Models, inside your agent. On Hermes Agent you use the AI Model card on your server page.
- Keep your trial server. If you started with the free trial, pay for Starter before the 7 days end. Your server, your agent and everything on it stay as they are.
- Read the docs. Browse the rest of this documentation to learn about advanced features like Setting Up Your AI Provider, Channel Management, and server monitoring.
Tip
Bookmark your dashboard URL. You can check on your server, view metrics, open your agent and manage your plan from any device with a web browser.