Fetchlake
← Blog
Research

Hermes Agent: The Easiest Way to Run Your Own Autonomous AI — At Almost No Cost

Matthew White · 1 June 2026
Hermes Agent: The Easiest Way to Run Your Own Autonomous AI — At Almost No Cost

If you already pay for an OpenAI subscription, you're closer than you think to running a fully autonomous AI agent that can browse the web, send emails, remember things across sessions, and work while you sleep. The Hermes agent from Nous Research changes what's possible for everyday users — and the setup takes less than a minute.

Why Autonomous Agents Are a Game Changer

There's a fundamental difference between a chatbot and an agent.

A chatbot answers your questions. An agent acts on them.

When you ask a chatbot to research a topic, it gives you text. When you ask an agent to do the same thing, it opens a browser, reads the relevant pages, pulls together a summary, cross-references sources, and emails it to you — all without you lifting a finger again.

This is what makes Hermes significant. We're not talking about a marginally better assistant. We're talking about a shift from AI that responds to AI that operates. The agent runs a closed learning loop: it creates skills, recalls memory across sessions, and refines its own capabilities over time. It isn't just executing a single command — it's building a working model of how to do things for you.

For businesses, researchers, writers, and developers, this kind of automation used to require either expensive enterprise software or deep technical knowledge to build from scratch. Hermes brings it to anyone with a laptop — or a $6/month server.

What You Need Before You Start

The good news: if you already have an OpenAI subscription, you have most of what you need.

Hermes supports multiple model providers including OpenAI, OpenRouter, and the Nous Portal (their own bundled subscription). You can point it directly at your existing OpenAI API key and it works immediately. No new AI subscription required.

Beyond that, you need:

  • A computer running Linux, macOS, Windows (via WSL2 or native PowerShell), or even Android via Termux
  • About 60 seconds

That's it.

Installing on Your Home Computer First

Before we get to the server, it's worth getting familiar with Hermes on your own machine. The install is a single command.

Linux, macOS, or WSL2:

curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash

Windows (PowerShell):

iex (irm https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.ps1)

Android (Termux): The Linux installer works here too — it auto-detects the environment.

Once installed, run the setup wizard:

hermes setup --portal

This walks you through OAuth authentication and connects the four core Tool Gateway features: web search, image generation, text-to-speech, and browser automation. You're up and running with access to 60+ built-in tools.

Running locally is great for testing. The limitation is that when your computer is off, the agent stops. For anything you want running around the clock — scheduled tasks, email monitoring, automated research pipelines — you want it on a server.

Why Host on a VPS?

A VPS (Virtual Private Server) is a small cloud computer that runs 24/7. You rent it by the month, it sits in a data centre, and your agent keeps working regardless of what you're doing on your own machine.

The benefits beyond uptime:

  • Always on. Your agent can monitor emails, run schedules, and respond to triggers at 3am without your laptop being open.
  • Security. Your API keys, credentials, and agent memory never sit on your personal device. If your laptop is lost or handed to someone, none of that is on it.
  • Stable connection. A data centre connection is faster and more reliable than home broadband for anything involving web scraping or API calls.

Setting Up Your VPS on Hostinger

I recommend Hostinger for this — their VPS plans are well priced, the control panel is beginner-friendly, and their support is responsive.

For running Hermes, the KVM 1 plan at $6.49/month is more than enough to start. You get:

  • 1 vCPU core
  • 4 GB RAM
  • 50 GB NVMe storage
  • 4 TB bandwidth

That's serious headroom for an AI agent. If you want to run multiple agents or heavier workloads down the line, the KVM 2 plan at $8.99/month gives you 2 cores and 8 GB RAM.

Step 1 — Sign up and create your VPS

  1. Head to Hostinger and select a VPS plan.
  2. During setup, choose Ubuntu 22.04 as your operating system — it's the most straightforward for this kind of work.
  3. Set a strong root password, or better yet, choose the SSH key option.
  4. Choose a data centre region close to you.
  5. Once provisioned (usually under 2 minutes), you'll see your server's IP address in the Hostinger dashboard.

Step 2 — Connect to your server

Open a terminal on your computer and connect via SSH:

ssh root@YOUR_SERVER_IP

Replace YOUR_SERVER_IP with the IP shown in your Hostinger dashboard. Type your password when prompted. You're now inside your server.

Step 3 — Install Hermes

Run the one-line installer:

curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash

Then run setup:

hermes setup --portal

Follow the prompts — this is where you connect your OpenAI key or Nous Portal account and enable tools.

Step 4 — Keep the agent running after you disconnect

When you close your SSH session, any running process stops. To keep Hermes running permanently, install it as a background service.

Create a systemd service file:

nano /etc/systemd/system/hermes.service

Paste in the following:

[Unit]
Description=Hermes Agent
After=network.target

[Service]
ExecStart=/usr/local/bin/hermes start
Restart=always
RestartSec=5
User=root

[Install]
WantedBy=multi-user.target

Save with Ctrl+O, then Enter, then Ctrl+X to exit. Now enable and start it:

systemctl daemon-reload
systemctl enable hermes
systemctl start hermes

Your agent now starts automatically whenever the server reboots and restarts itself if anything goes wrong.

Using VS Code to Chat With Your Agent

This is where it gets genuinely nice. Instead of going back and forth in a terminal window, you can use Visual Studio Code as a comfortable, persistent interface for managing and chatting with your Hermes agent — all from your own computer, with the agent running on your Hostinger server.

Step 1 — Install VS Code

Download and install VS Code on your computer if you don't already have it.

Step 2 — Install the Remote-SSH extension

  1. Open VS Code.
  2. Click the Extensions icon in the left sidebar (or press Ctrl+Shift+X).
  3. Search for Remote - SSH.
  4. Install the one published by Microsoft.

Step 3 — Connect to your Hostinger server

  1. Press Ctrl+Shift+P (or Cmd+Shift+P on Mac) to open the command palette.
  2. Type Remote-SSH: Connect to Host and select it.
  3. Enter your server address:
root@YOUR_SERVER_IP
  1. VS Code opens a new window connected directly to your server. You'll see SSH: your-server-ip in the bottom-left corner confirming the connection.

Step 4 — Open a terminal and start chatting

  1. In VS Code, go to Terminal → New Terminal (or press Ctrl+`).
  2. This terminal is running on your server, not your local machine.
  3. Check your agent is running:
systemctl status hermes
  1. Open the Hermes chat interface:
hermes chat

You're now chatting with your agent directly from VS Code. The interface feels like any other terminal conversation, but it's running on a server that never switches off.

Why VS Code works so well for this

Beyond just chatting, VS Code gives you the full picture in one place:

  • File browser — browse your agent's memory files, logs, and config in the left sidebar
  • Multiple terminals — open one tab for chat, another for logs (journalctl -fu hermes), another for config edits
  • Syntax highlighting — when you edit config files or agent skill definitions, you get proper highlighting

It turns what would otherwise be a fairly bare terminal experience into something much more workable.

Giving Your Agent Access to Gmail

One of Hermes's most powerful features is email integration. The agent supports 27+ messaging platforms — email included — meaning you can wire it into your Gmail account and have it send messages, respond to queries, or trigger workflows based on what lands in your inbox.

Step 1 — Create a Google Cloud project

  1. Go to console.cloud.google.com and sign in.
  2. Click Select a project at the top → New Project.
  3. Give it a name (e.g. "Hermes Agent") and click Create.

Step 2 — Enable the Gmail API

  1. In the left menu, go to APIs & Services → Library.
  2. Search for Gmail API and click it.
  3. Click Enable.

Step 3 — Create OAuth credentials

  1. Go to APIs & Services → Credentials.
  2. Click Create Credentials → OAuth client ID.
  3. If prompted, configure the consent screen first — choose External, fill in your app name and email, and save.
  4. For application type, choose Desktop app.
  5. Click Create and download the credentials JSON file.

Step 4 — Transfer credentials to your server

From your local machine, copy the credentials file to your Hostinger VPS:

scp ~/Downloads/credentials.json root@YOUR_SERVER_IP:/root/.hermes/gmail_credentials.json

Then in your Hermes config, point the email integration at that file and run the OAuth flow. Hermes will give you a URL to open in your browser — log in with your Google account, grant permission, and that's it. Hermes stores a refresh token and can send and read email on your behalf from that point forward.

What your agent can now do with email:

  • Send you daily summaries of completed tasks
  • Email contacts on your behalf when triggered by a workflow
  • Monitor your inbox and act on flagged messages
  • Draft emails and queue them for your review before sending

The security model is worth noting: Google's OAuth means Hermes never sees your password. You're granting scoped access that you can revoke from your Google account at any time.

The Bottom Line

A year ago, setting up an autonomous AI agent meant either paying for an enterprise platform or spending weeks stitching together unreliable tools. Today, Hermes installs in 60 seconds, runs on a Hostinger VPS that costs less per month than a takeaway coffee, connects to your email and the web out of the box, and you can manage the whole thing from VS Code like a pro.

If you already have an OpenAI subscription, the marginal cost to get started is basically nothing. The gap between "I wish AI could just handle this for me" and actually having an agent that does has never been smaller.

Hermes is developed by Nous Research. This article contains a referral link to Hostinger — I recommend them because they're genuinely good value, and using the link supports this content at no extra cost to you.