How to install Hermes Agent on a VPS

Hermes Agent is an autonomous AI agent that can run not only on a local laptop, but also on a remote server. This is useful when you need long-running jobs, Telegram access, integrations, and a stable automation environment that does not depend on one workstation. This guide explains how to install Hermes Agent on an Ubuntu VPS, complete the basic setup, and prepare the server for real operating use.

Contents

• Why install Hermes Agent on a VPS
• What you need before installation
• Step 1. Prepare the server
• Step 2. Install Hermes Agent
• Step 3. Run the first setup
• Step 4. Check that Hermes Agent works
• Step 5. Decide how you will access the agent
• Step 6. Prepare the agent for background work
• Step 7. Do not forget security
• What to do if something goes wrong
• Conclusion
• Where to run Hermes Agent

Why install Hermes Agent on a VPS

If the agent runs only on a local computer, it depends on that computer: the laptop can be turned off, the network connection can drop, and background jobs can stop. A VPS removes these limitations.

Main advantages:

• the agent can stay available from the network;
• cron jobs and integrations can run continuously;
• Telegram and other communication channels are easier to connect;
• your personal workstation and automation infrastructure are separated;
• resources can be scaled as the workload grows.

What you need before installation

For a basic setup you need:

• a VPS with Ubuntu 24.04 or Ubuntu 22.04;
• SSH access;
• a user with sudo privileges;
• internet access for downloading the installer;
• an account and model provider that Hermes Agent will use.

For a first launch without heavy workloads, a typical starter server is enough:

• 2 vCPU;
• 4 GB RAM;
• 20–40 GB SSD/NVMe;
• stable network access.

If you plan to run many tasks, cron jobs, a Telegram gateway, and several integrations, it is better to keep some reserve in RAM and disk space from the beginning.

Step 1. Prepare the server

Connect to the server:

ssh user@your_server_ip

Update the packages:

sudo apt update && sudo apt upgrade -y

Install the common utilities needed for CLI tools and administration:

sudo apt install -y curl git unzip build-essential

If you are working as root, create a separate user for long-term operation instead of keeping the whole agent environment in the root session.

Step 2. Install Hermes Agent

For Linux, macOS, WSL2, and other Unix-like systems, Hermes Agent documentation recommends the shell installer:

curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash

This method installs the CLI and prepares the working environment without manual build steps.

After installation, check that the command is available:

hermes --help

If the command is not found, open a new SSH session or reload the current user’s shell profile.

Step 3. Run the first setup

After installation, run the basic setup. The quick path recommended in the Hermes Agent documentation is:

hermes setup --portal

This mode is convenient if you want to connect a model and the initial configuration through Nous Portal.

If you prefer a step-by-step setup, use:

hermes setup
hermes model
hermes doctor

At this stage, check that you:

• selected the model provider;
• verified that Hermes sees its configuration and environment;
• confirmed that the basic dependencies are installed correctly;
• decided whether you need only the CLI or also Telegram/gateway access.

Step 4. Check that Hermes Agent works

For a quick check, run a one-off request:

hermes chat -q "Hello! Briefly describe what you can do."

To open an interactive session:

hermes

If Hermes responds and the CLI works without errors, the basic installation is complete.

Step 5. Decide how you will access the agent

One of the strongest Hermes Agent scenarios is access not only through a terminal, but also through messengers. This is especially useful on a VPS because the agent lives on the server and does not depend on a local machine.

A typical path looks like this:

1. install Hermes Agent on a VPS;
2. connect a model and basic configuration;
3. configure the gateway;
4. connect Telegram or another channel;
5. use the agent remotely for tasks, research, automation, and background workflows.

If you need the Telegram scenario specifically, configure the gateway and platform parameters after the basic setup.

Step 6. Prepare the agent for background work

One reason to move Hermes to a VPS is the ability to keep it as a permanently available operating layer, not a one-off local CLI.

Useful next directions:

• scheduled cron jobs;
• research and monitoring;
• autonomous content workflows;
• processing incoming Telegram messages;
• long-running integrations with external services.

In practice, think through:

• where the configuration will be stored;
• which secrets the agent needs;
• which tasks will run on schedule;
• whether you need a separate user and a separate working directory;
• whether configuration and log backups are required.

Step 7. Do not forget security

Even for a first test setup, it is better not to leave the server in a raw state.

Minimum baseline:

• SSH key access instead of simple passwords;
• an updated system;
• limited open ports through a firewall;
• a separate user for application work;
• careful storage of API keys and tokens;
• a clear understanding of who can access the agent and through which channel.

If the server is used for Telegram, cron, and external integrations, security becomes even more important: this is already a persistent networked environment, not just an experimental CLI session.

What to do if something goes wrong

Below are several common checks.

The hermes command is not found

Check whether:

• the installation finished without errors;
• PATH was picked up in a new shell session;
• you are using the same user that ran the installer.

Initial setup fails

Check whether:

• the server has outbound internet access;
• the selected model provider is available;
• authentication parameters are not missing;
• hermes doctor shows any actionable errors.

The agent works locally but is not ready for permanent operation

This usually means the basic installation is done, but several operating questions are still open:

• secrets;
• background job design;
• gateway setup;
• updates;
• firewall and external access.

In that case, treat server preparation as a separate next step instead of assuming that the initial installation is already production-ready.

Conclusion

Installing Hermes Agent on a VPS is a practical start for scenarios where an AI agent needs to stay available continuously, not only while a local terminal is open. It is useful for automation, remote access through Telegram, background jobs, and a more stable operating model.

The installation itself is straightforward: prepare an Ubuntu server, run the official installer, complete setup, and check the first command. After that, you can move on to integrations, gateway configuration, cron jobs, and workflows for your own processes.

Where to run Hermes Agent

Hermes Agent is especially useful when you need persistent access: background workflows, cron jobs, research, integrations, and Telegram. For these tasks, VPS/VDS hosting is more convenient than a local machine because the agent does not depend on a laptop, remains reachable over the network, and can be scaled as workload grows.

If you want to run Hermes Agent not as a test, but as a working server-side setup, the logical next step is to choose a VPS for your scenario.