How to Get Started with Managed AI Agents on Cloudways

Cloudways Managed AI Agents allow you to deploy, configure, and manage AI-powered agents on managed cloud infrastructure without handling the underlying server setup yourself.

These agents can help automate tasks, interact with supported communication channels, connect with AI models and external services, and perform agent-specific workflows based on their capabilities.

Cloudways supports multiple Managed AI Agents, including OpenClaw and Hermes. Each agent is designed for different use cases and may offer its own features, integrations, communication channels, AI model support, and configuration options.

With Managed AI Agents on Cloudways, you can select an available agent, choose a suitable server or instance size, complete the required configuration, and manage the agent directly from the Cloudways Platform.

This guide explains how to get started with Managed AI Agents on Cloudways, including how to choose an available AI agent, deploy it, configure the required settings, access its management interface, and understand the options available after deployment.

Managed AI Agents is a Cloudways product line that helps users deploy AI agents on dedicated infrastructure without managing the technical setup themselves.

In a traditional self-hosted setup, users may need to manually configure a server, install Docker, set up SSL certificates, manage updates, and troubleshoot the runtime environment. These steps can be difficult for non-technical users and time-consuming even for experienced developers.

Cloudways simplifies this process by managing the infrastructure required to run the agent. You choose the agent, select the required instance size and region, and launch it from the Cloudways Platform.

Cloudways handles the server provisioning and agent deployment, while you focus on configuring and using the agent.

OpenClaw is an AI agent you run on your own machine, working with whichever model you prefer, whether that is Claude, GPT, Gemini, or a local one, using your own API keys. Everything runs through a single always on service called the gateway, which handles your messaging apps and manages memory and model calls. That means you just message the agent in WhatsApp, Telegram, Slack, or Signal and it picks up where you left off. It can run shell commands, read and write files, browse the web, check your inbox and calendar, monitor builds, run scheduled jobs, and send notifications to your phone.

You add new abilities as skills rather than editing code, with over 100 ready made ones available, and it can spin up extra agents to work on several things at once. Its strength is being a broad assistant that lives in the apps you already use and can run a whole team of agents.

Hermes Agent is the main alternative, built by Nous Research around one idea, that a single agent should get better at your work the more you use it. You mostly use it from the command line, with messaging as an optional extra. When it works out how to do something tricky, it saves the method as a reusable skill it can later refine on its own, and it keeps notes on your preferences, projects, and setup alongside a searchable history of past sessions. It ships with a wide toolset covering web search, files, terminal, browsing, images, speech, and code execution, and it is unusually flexible about where that code runs, offering local, Docker, SSH, and several cloud sandbox options.

It can also hand work to child agents with limited permissions or run tasks on a schedule you describe in plain English. Its strength is being one agent that steadily learns your workflows, with strong options for developers.

When launching Managed AI Agents, you need to choose an instance size based on your expected workload. A workload means the amount of activity your agent is expected to handle. A simple workflow with low usage may need fewer resources, while a more complex workflow or higher traffic may require more CPU, memory, storage, and bandwidth.

The available instance sizes are:

Before launching a Managed AI Agent on Cloudways, make sure you have an active paid Cloudways account.

Managed AI Agents are not available on trial accounts, so you need to upgrade to a paid account before you can access and deploy them from the Cloudways Platform.

You may also need an API key or other credentials from a supported Large Language Model (LLM) provider, such as OpenAI, Anthropic, Google, or another provider supported by your selected agent.

An API key is a secure credential that allows the Managed AI Agent to connect to the AI model provider and use the selected model. Depending on the agent and configuration, you can add these credentials during or after the setup process.

Follow these steps to launch and configure your managed AI agent from the Cloudways Platform.

A region represents the physical data center location where your agent runs. It is recommended to select a region based on your data residency or compliance requirements.

Choose an instance size based on your expected workload and agent requirements.

The available instance sizes are:

For basic usage or testing, the lowest instance size may be sufficient. For complex workflows, multiple tasks, or higher usage, choose an instance with more CPU and memory resources.

An LLM (Large Language Model) provider powers your OpenClaw agent by providing the AI model used to process instructions and generate responses.

An API key is a secure authentication key that allows OpenClaw to connect with your LLM provider.

If you do not have an API key, click Get API Key to create one from your selected provider.

Once your OpenClaw agent is successfully created, it will appear under the Managed AI Agents section.

You can perform the following actions:

After launching your OpenClaw agent, you can manage your LLM provider connection from the OpenClaw dashboard.

The LLM provider connection allows your agent to use the selected AI model for processing requests and completing tasks.

If you have not added your LLM Key in the first step, you can also update your API key anytime from LLM Provider section inside AI Agents tab.

After your OpenClaw agent is deployed, you can access the dashboard directly from the Cloudways Platform.

The OpenClaw dashboard allows you to:

The dashboard can be accessed using the Open Agent option from the Managed AI Agents section.

To learn How to Manually Pair Your Device with a Managed AI Agent on Cloudways, click here.

That’s it! We hope this article was helpful.