斜杠中年斜杠中年AI × 沟通 × 商业 × 人生
AI Practical Guide

Self-Host Hermes AI Agent on a Hostinger VPS: The Complete Install Guide

Hermes Agent is an open-source AI agent that runs in your terminal and on Telegram, WhatsApp, and a dozen other platforms. Here's how to install it on a Hostinger VPS so it runs 24/7, keeps its own memory and skills, and never shuts down when you close your laptop.

2026-09-05Updated: 2026-09-0511 min readWesley Chong
#hermes agent#self-hosted ai#hostinger vps#server setup#ai agents#open source ai#linux
Self-Host Hermes AI Agent on a Hostinger VPS: The Complete Install Guide|AI Practical Guide 封面图

Summary

Hermes Agent is an open-source AI agent that runs in your terminal and on Telegram, WhatsApp, and a dozen other platforms. Here's how to install it on a Hostinger VPS so it runs 24/7, keeps its own memory and skills, and never shuts down when you close your laptop.

One-Sentence Answer

Hermes Agent is an open-source AI agent that runs in your terminal and on Telegram, WhatsApp, and a dozen other messaging platforms — and you can self-host it on a cheap Hostinger VPS so it runs 24/7, keeps its own memory and skills, and never shuts down when you close your laptop.

If you have only seen Hermes as something that lives on your own PC, this guide walks you through putting it on a server where it can stay online permanently. First, the explainer on what Hermes actually is.

The Hostinger Gotcha (Read This First)

Before you run a single command, confirm you actually have a VPS — because Hermes needs a shell and root (or sudo) access, and not every Hostinger product has that:

| Hostinger plan | Can it run Hermes? | | --- | --- | | VPS (KVM) | ✅ Yes — full root SSH access | | Shared hosting | ❌ No — managed, no root SSH | | Cloud hosting | ❌ No — managed, no root SSH |

Grab your credentials in hPanel → VPS → SSH access (server IP + root password). If you bought a VPS, you will see an SSH section there. If you are on Shared or Cloud, you will need to upgrade first.

What You Need Before You Start

  • A Hostinger VPS (KVM) running Ubuntu 22.04 or 24.04 LTS.
  • SSH access with root, or a user that can sudo.
  • curl and xz-utils — the installer downloads Node.js as a .tar.xz archive.
  • An LLM provider — a Nous Portal OAuth login, an OpenRouter/Anthropic/DeepSeek API key, or any of the 20+ supported providers.
  • A domain or static IP only if you plan to expose the web dashboard, webhooks, or the API server publicly.

Step-by-Step Install

1. SSH in and update

ssh root@<your-vps-ip>
sudo apt update && sudo apt upgrade -y

2. Install the prerequisites

sudo apt install -y curl xz-utils git ca-certificates

Run Hermes under its own user rather than root — it keeps things tidy and makes the service easier to reason about:

sudo adduser hermes
sudo usermod -aG sudo hermes
su - hermes

4. Run the official installer

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

The installer detects what is missing (uv, Python, Node.js) and installs it for you. On Linux, curl and xz-utils are the only things you are expected to have pre-installed.

5. Reload your shell and verify

source ~/.bashrc
command -v hermes     # should print a path, e.g. /home/hermes/.local/bin/hermes

6. Run the setup wizard

hermes setup          # interactive: model, provider, terminal, gateway, tools
hermes doctor         # health check — confirm no missing dependencies

hermes setup walks you through picking a model and provider. Use hermes auth add <provider> (for example hermes auth add nous for the OAuth portal login, or hermes auth add openrouter) to attach credentials without pasting keys into config files.

7. First smoke test

hermes chat -q "Reply with the single word: online"

If you get a sensible reply, the agent is working on the server.

Keep It Running 24/7 (The Gateway)

A one-off hermes chat dies the moment you close your SSH session. To make Hermes always-on — connected to Telegram or WhatsApp and running scheduled jobs — you need the gateway as a service:

hermes gateway setup    # pick and configure your messaging platforms (Telegram, etc.)
hermes gateway install  # register as a systemd background service
hermes gateway start
hermes gateway status

hermes gateway install handles the systemd wiring for you. Two things to remember:

  • Start the gateway before installing the service. Run hermes gateway setup (or hermes gateway run once to confirm it works) so the service has a valid config to launch.
  • If you manage systemd yourself, use the exact path from command -v hermes, and run sudo loginctl enable-linger $USER so the process survives logout on SSH-only boxes.

Once it is running, the same agent answers you on Telegram from the server — with memory, skills, cron jobs, and full terminal access, none of which depend on your laptop staying open.

What You Can Do Once It Is Up

| Capability | Command | | --- | --- | | Interactive chat | hermes | | Single query | hermes chat -q "..." | | Run as a messaging bot | hermes gateway run | | Scheduled jobs | hermes cron create "0 9 * * *" | | Multiple isolated agents | hermes profile create <name> | | Web dashboard | hermes dashboard | | OpenAI-compatible proxy | hermes proxy |

A natural next step after the install: have the agent watch over the very VPS it lives on, or expand it into a whole team of specialist bots.

Troubleshooting

| Problem | Likely cause | Fix | | --- | --- | --- | | hermes: command not found | Shell path not reloaded | source ~/.bashrc, reopen SSH, or check command -v hermes | | Installer fails | Missing curl/xz-utils/git | sudo apt install -y curl xz-utils git ca-certificates, then retry | | hermes doctor flags missing deps | Incomplete install | Re-run the installer and review the output | | Gateway service exits immediately | Gateway not configured | Run hermes gateway setup first, then hermes gateway install | | Agent stops after SSH logout | Running interactively, not as a service | Use hermes gateway install or sudo loginctl enable-linger $USER | | Provider auth fails | Credentials not set | Run hermes auth add <provider> or hermes setup again | | Server runs out of memory | Local model too large | Switch to a hosted model, or upgrade RAM |

Key Takeaways

  • Only a VPS works. Hostinger VPS (KVM) has root SSH; Shared and Cloud hosting do not.
  • Install is one command. curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash, then hermes setup.
  • Verify before you trust it. command -v hermes and hermes doctor confirm the install is clean.
  • Always-on needs a service. The gateway runs as systemd via hermes gateway install — otherwise it dies with your SSH session.
  • The agent keeps its own memory and skills on the server, so it improves the longer it runs.

To understand what makes Hermes different from a plain chatbot, start with the explainer, then see how it keeps a VPS healthy on its own.

FAQs

Can I install Hermes on Hostinger Shared or Cloud hosting?

No. Hermes needs a real shell with root or sudo access, which only Hostinger VPS (KVM) plans provide. Shared and Cloud hosting are managed, so you cannot SSH in with root. If you are unsure which plan you have, check hPanel under VPS for the SSH access section.

Do I need a powerful VPS to run Hermes?

Not really. The agent itself is lightweight, and the heavy lifting happens on the model provider's side (OpenRouter, Anthropic, DeepSeek, etc.) when you use a hosted model. A small VPS with 1-2 GB of RAM is usually enough unless you also run a local model on the same box.

Will the agent keep running after I close my SSH session?

Only if you run it as a service. Interactive sessions die when you disconnect. Use `hermes gateway install` to register it as a systemd background service, or `sudo loginctl enable-linger $USER` so your processes survive logout.

Which LLM provider should I use on the VPS?

Any of the 20+ supported providers work. The quickest path is a Nous Portal OAuth login (`hermes auth add nous`) or an OpenRouter API key. You can also point at DeepSeek, Anthropic, OpenAI, Google Gemini, or a local model if the box has enough RAM.

分享这篇文章 / Share Article
Wesley Chong

Author

Wesley Chong

Software developer, digital consultant, and Toastmasters speaker from Kluang, Malaysia.

Focusing on helping ordinary people upgrade communication, expression, business, and life with AI.

Related Reading

Build Your AI Team with Hermes Bot Mode: Turn Every Profile into a Dedicated Bot|AI Practical Guide 封面图
AI Practical Guidehermes agentai agents

Build Your AI Team with Hermes Bot Mode: Turn Every Profile into a Dedicated Bot

Hermes Agent's Bot Mode turns every Hermes profile into a named bot with its own role, model, memory, skills, and avatar. Bots run recurring routines, deliberate together in group chats, and message each other directly — even across machines. There is no new primitive to learn: a bot is a profile, so everything is visible from the command line too.

2026-08-2711 min read