Run Hermes on a VPS (step-by-step)

Running an AI agent on your laptop works right up until you close it. This guide moves it somewhere that never closes.

1. What is Hermes Agent?

Hermes Agent, usually just called Hermes, is an open-source AI agent from Nous Research. It can use tools, remember information and work automatically in the background. Running it on a VPS means it stays online 24/7, even when your computer is off.

You do not need a server just to try it. Hermes Agent has a desktop app for Apple Silicon Macs and for Windows, from the Hermes website. This guide is for when you want it running while your own computer is asleep.

If either “agent” or “VPS” is unfamiliar, read what is an AI agent and what is a VPS first. They are five minutes each and the rest of this makes much more sense afterwards.

2. What you’ll need

• A VPS running Ubuntu. The cheapest tier is enough. Hermes Agent’s docs test on the latest Ubuntu, and say most Linux systems with glibc and systemd work too.
• A terminal on your own computer. Mac and Linux have one built in; on Windows use PowerShell.
• An API key for whichever model you want the agent to use.

3. Set up your VPS

Create the server, choose Ubuntu, choose the smallest size, and pick a location near you. You will be given an IP address and login details.

Connect to it:

ssh root@YOUR_SERVER_IP

The first thing to do on any new server is update it:

apt update && apt upgrade -y

4. Install Hermes Agent

The script needs git, curl and tar. Ubuntu usually has all three; if not, apt install -y git curl tar adds them. Then run Hermes Agent’s own install script, the one its docs give for Linux, macOS and WSL2:

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

It downloads everything Hermes Agent needs (its own Python and Node.js included), creates the hermes command and, when you run it by hand like this, offers to walk you through setup. When it finishes, reload your shell with source ~/.bashrc.

Read the script (open the URL in a browser, or run curl without the pipe to bash first) before you run it on a server you care about. That is good practice for any installer, not just this one.

Other official ways in, for reference: a native Windows installer run from PowerShell, desktop apps for Apple Silicon Macs and Windows, and an official Docker image. A Docker install does not use hermes update; you update it by pulling a new image. The installation docs list all of them.

5. Connect it to a model

Hermes Agent’s documented quick setup is:

hermes setup --portal

This signs you in to Nous Portal, Nous Research’s paid subscription, sets it as the model provider and switches on its tools, rather than pasting a raw API key into a config file. If you would rather use an API key from another provider, run hermes model instead and pick it there. On a VPS with no browser, that means opening the URL it gives you on your own laptop or phone to finish the sign-in, then returning to the terminal. We have not run this step on a live VPS ourselves yet, so treat the exact prompts as not yet verified on a live server: follow what the command actually shows you over what is written here if the two disagree.

6. Keep it running after you disconnect

Closing your terminal normally stops whatever was running in it. The simplest fix, with no Hermes-specific commands, is to start it inside a tmux session (a terminal that keeps running):

tmux new -s hermes
# run Hermes here

Press Ctrl+b then d to detach; the session and whatever is running inside it keep going. Reattach any time with tmux attach -t hermes.

If you want it to survive a full server reboot as well, Hermes Agent’s docs now cover that. Its gateway (the part that keeps running and answers you in a chat app such as Telegram or Discord) installs itself as a service. Set up the chat app first with hermes gateway setup, then, for a VPS that should come back on its own at boot:

sudo hermes gateway install --system
sudo hermes gateway start --system
sudo hermes gateway status --system

The docs recommend this system service for a VPS. These commands come from the docs and are not yet verified on a live server by us.

7. Configure memory

With the install script, Hermes Agent keeps its settings, memory and skills in a .hermes folder in your home directory (~/.hermes/). A Docker install keeps them in whichever folder you mount into the container. Wherever it lives, the same advice applies: keep it small and specific. A memory full of everything is as useless as no memory at all, and it costs more, because all of it gets sent to the model each time.

8. Test it works

Give it a small job with an obvious right answer, something you can verify in a second. If that works, give it something slightly harder. Reattach to the tmux session to see its output at any time.

9. Useful tips

• Set a spending limit on your model provider account before leaving it running. An agent in a loop is an expensive agent.
• Use a cheap model for the agent itself. Most of what a scheduled agent does is wake up, check whether anything needs doing and decide that it does not. That work does not need a clever model, it needs a cheap one. Keep the expensive models for the jobs that genuinely need thinking, and have the agent call out to them only then. You are the one deciding where that line sits, and putting it too low gives you an agent that is confidently wrong for a week before you notice.
• Give it the narrowest access that works. One folder, not your whole server.
• Check on it weekly at first. You will find out quickly whether it is doing anything useful or quietly failing.
• Write down what you changed. In three months you will not remember.

10. Next steps

Once it is running, the interesting question is what to give it. Start with one small, repetitive job you actually do, not the most impressive thing you can imagine. Boring and working beats ambitious and broken.

If you would rather have help setting this up for your own situation, that is what the work with me page is for.