OpenClaw

OpenClaw is an open solution for deploying and managing AI assistants that allows you to run your own LLM services, work with LLM models, and integrate them into your projects without complex infrastructure.

The platform is suitable for creating chatbots, internal AI assistants, text processing services, data analysis, and other scenarios that require flexible work with language models.

You can create a VPS with OpenClaw in the Beget catalog of ready-made applications to quickly deploy a ready-made environment for working with AI. OpenClaw allows you to connect various models, manage configurations, customize access, and use APIs to integrate with websites, CRM systems, Telegram bots, and internal services.

The platform can be installed in just a few clicks—you don't need to manually configure the environment, web server, or SSL. The rental cost depends only on the selected server configuration, which allows you to flexibly scale resources to the load: from test benches to production scenarios.

OpenClaw from Beget is a ready-to-use build with your or a free domain and an installed SSL certificate, which can be deployed in minutes and is immediately ready for integration into your projects.

Installation package information

• Ubuntu 24.04
• Nginx
• Node.js
• OpenClaw

OpenClaw installation

To install OpenClaw along with creating a server, you will need to specify the following additional parameters:

• The domain name under which the OpenClaw platform will be accessible. You can either choose your own domain or register a new one. In addition, you can use a free domain in the .beget.app zone. A free SSL certificate will be installed on the specified domain.

In the “Additional OpenClaw parameters” block, you can also specify:

• Login for the web interface. Used to access the OpenClaw web interface.
• Password for the web interface. Used to access the OpenClaw web interface.
• Administrator email.

After creating the virtual server, we will automatically deploy and configure OpenClaw. It usually takes up to 5 minutes. You will receive a notification once the deployment is complete.

Getting started

To get started, open the OpenClaw information window and follow the link for your first login, which is available under the “Application Data” block. It looks like this: https://my-domain.beget.app?token=<token>, where my-domain.beget.app is the domain specified during installation or a free .beget.app domain, and <token> is the access token.

To log in, use the username and password for the web interface that you specified when creating the virtual server. These were also sent to the contact email address for the account and are available in the information window for the installed application in the virtual server control panel.

After authorization, you will be taken to the OpenClaw dashboard.

To work with OpenClaw via the web interface, you need to configure the pairing of the device you are using to log in. It must be done each time you log in from a new browser or device.

To do this:

1. Connect to the server via SSH.
2. Switch to the OpenClaw user with the command: su - openclaw.
3. Get a list of pairing requests with the command: openclaw devices list.
   It will look like this:

openclaw@zucplcjiyt:~$ openclaw devices list

🦞 OpenClaw 2026.3.2 (85377a2) — Say "stop" and I'll stop—say "ship" and we'll both learn a lesson.

│
gateway connect failed: Error: pairing required
◇
Direct scope access failed; using local fallback.
Pending (2)
┌──────────────────────────────────────┬───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┬──────────┬───────────────┬──────────┬────────┐
│ Request                              │ Device                                                                                                                                                                                        │ Role     │ IP            │ Age      │ Flags  │
├──────────────────────────────────────┼───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┼──────────┼───────────────┼──────────┼────────┤
│ da402295-4177-4cd4-afba-3a70e972c5c3 │ c37e69a017863f6d9be128e3bd98ceb8418c121472a701c5401e8ee52a06522c                                                                                                                              │ operator │               │ just now │        │
│ f76c56f0-7f26-4512-ac03-0e9df5128d1a │ ded709f40c00f50644ef58547d0d3b6e2f0ecdb1ad8bac06da1c53eff096ea8a                                                                                                                              │ operator │ 94.19.120.239 │ just now │        │
└──────────────────────────────────────┴───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┴──────────┴───────────────┴──────────┴────────┘

4. Copy your authorization request ID, which can be found in the Request column of the table. If there are several authorization requests, you need to select the one with the IP address of the device you used to access the web interface in the IP column.
5. Confirm the device pairing with the command: openclaw devices approve <request-id>
   For example:

openclaw devices approve f76c56f0-7f26-4512-ac03-0e9df5128d1a

6. Refresh the web interface page.

The device pairing is now configured, and you can use the OpenClaw web interface.

Connecting models

To connect LLM models, such as OpenAI, follow these steps:

1. Connect to the server via SSH.
2. Switch to the openclaw user by running the command: su - openclaw
3. Start the initial setup process: openclaw onboard
4. For the first three questions of the setup wizard, select the following options:
   1. Yes
   2. QuickStart
   3. Use existing values
5. Next, select a model provider, such as OpenAI.
6. Specify the authentication method. There are two options available for OpenAI:
   1. authorization by API key;
   2. authorization via OAuth.
7. Enter the API key or complete authorization in the browser, depending on the method you selected.
8. Select one of the suggested models.
9. After that, you can set up a channel in OpenClaw for communicating with the assistant, such as a Telegram bot. You can skip this step, in which case interaction will be available via the web interface.
10. Complete the adaptation process by configuring additional parameters or leaving the default values.

After adding the model, go to the chat and check that it is working.

To add other models or configure channels, simply run the process again.

FAQ