Claude Code is an increasingly popular tool for developers who want to use artificial intelligence to write code faster, manage projects, implement automation and build their own AI agents. If you use it more seriously, it quickly becomes apparent that a VPS server is a far more practical choice than a local computer.

Most Claude Code users initially run it locally on their computer. This is perfectly adequate for testing, but limitations quickly become apparent when working for longer periods:

• the computer must be switched on at all times,
• if the connection is lost, the session may be interrupted,
• AI tasks running for several hours are halted,
• working from mobile devices is not the most practical,
• the development environment is not accessible 24/7.

That is why a VPS server is an excellent solution. Claude Code can run in a constantly accessible server environment, and you can access it via an SSH connection, no matter where you are.

Why use a VPS for Claude Code

When using Claude Code on a VPS server, raw processing power isn’t the most important factor. AI processing is mostly carried out within Anthropic’s infrastructure, whilst the VPS serves as a permanent development environment where you store your projects, repositories, configurations and automations.

This means that:

• the AI agent continues to work even when you close your laptop,
• projects and Git repositories remain accessible at all times,
• access the environment from anywhere,
• separate your development environment from your personal computer,
• run long-running automations and background processes.

In practice, a VPS can become your central AI workspace, where Claude Code helps manage code, Docker environments, test applications, cron jobs, MCP servers, databases and other development tools.

What kind of VPS do you need

For basic use of Claude Code, you do not need a very powerful server. As the main AI processing takes place in the cloud, the load on the VPS server is usually relatively low.

Usage	Recommended configuration
Testing and lighter projects	1 vCPU / 2 GB RAM
Serious development environment	2 vCPU / 4 GB RAM
Multiple agents, Docker, CI/CD	4+ vCPU / 8+ GB RAM

 

Why Ubuntu

Ubuntu is a very practical choice for Claude Code because it offers good support for development tools, Node.js, Docker and most of the libraries that developers use in AI automation.

We recommend one of the LTS versions:

• Ubuntu 24.04 LTS,
• Ubuntu 22.04 LTS.

Installing Claude Code on a VPS

Once the VPS server is ready, connect to it via SSH:

ssh root@IP_NASLOV

First, update the system:

apt update && apt upgrade -y

Then install the basic tools:

apt install -y curl git tmux

Installing Node.js

Claude Code requires a Node.js environment. We recommend using Node.js 22:

curl -fsSL https://deb.nodesource.com/setup_22.x | bash -
apt install -y nodejs

Check the installed version:

node -v
npm -v

Installing Claude Code

Install Claude Code via npm:

npm install -g @anthropic-ai/claude-code

After installation, run it with the command:

claude

On first launch, you will need to log in or authorise. The process will guide you to a link which you open in your browser.

Why use tmux

When working via an SSH connection, it is highly recommended to use tmux. This is a tool that allows the session on the server to remain active even when you disconnect.

Create a new tmux session:

tmux new -s claude

In this session, launch Claude Code:

claude

You can log out of the session using the following combination:

CTRL+B, nato D

Later, return to the same session with the command:

tmux attach -t claude

This way, Claude Code can continue working even if you close the terminal or lose your connection.

Access from a phone or tablet

One of the major advantages of a VPS environment is that you can access it from anywhere. On your phone or tablet, you can use SSH apps such as:

• Termius,
• Blink Shell,
• JuiceSSH.

This is particularly useful when you want to check the progress of an AI agent, confirm a change or quickly run an additional task.

Security recommendations

As this is a server environment, we recommend a few basic security measures.

Do not use the root user for day-to-day work

Create a separate user:

adduser claude
usermod -aG sudo claude

Use SSH keys

To access the server, we recommend using SSH keys and disabling password login wherever possible.

Limit open ports

In the basic configuration, you usually only need SSH access. You can configure the firewall using the following commands:

ufw allow OpenSSH
ufw enable

Update the system regularly

Perform updates regularly:

apt update && apt upgrade

What else can you run alongside Claude Code

As you have complete control over the environment on your VPS server, you can also combine Claude Code with other tools and services:

• Docker environments,
• PostgreSQL or MySQL databases,
• Redis servers,
• Git repositories,
• CI/CD processes,
• MCP integrations,
• local API services,
• cron automations,
• RAG systems,
• AI memory storage solutions.

In this way, VPS becomes a stable and always-available environment for developing AI automation.

Claude Code as an always-available AI agent

Claude Code on a VPS server is particularly useful when you want an AI assistant that isn’t tied to your local computer.

Typical use cases:

• automatic code refactoring,
• background development tasks,
• application monitoring,
• automated testing,
• documentation generation,
• AI support for internal processes,
• AI agent development,
• centralised AI workspace for the team.

VPS as a central AI workspace

If you’re using Claude Code more seriously, a VPS is a natural upgrade from a local development environment. It provides stable access, uninterrupted operation, better project organisation, and the possibility of advanced AI automation.

• For basic use, we recommend an Ubuntu VPS with at least 4 GB of RAM.
• For more demanding projects, Docker environments and multiple AI agents, a VPS with 8 GB of RAM or more is a better choice.

Note: Claude Code is a tool developed by Anthropic. To use it, you need appropriate access or a user account with the service provider.