Install Hermes Agent on Linux and Complete Your First Setup

Hermes Agent is an open-source AI agent that runs in your terminal and can also connect to messaging platforms, scheduled jobs, tools, and workflows.

This guide keeps the first run simple: install Hermes Agent on Linux, run the setup wizard, connect a provider, and confirm the CLI actually works before you move on.

When a Linux VPS is a good fit for Hermes Agent

A Linux VPS or server makes sense if you want Hermes Agent to:

  • run from a stable always-on machine

  • access local files or scripts on that server

  • handle scheduled jobs later

  • connect to things like Telegram, email, or other automation flows

A VPS is probably not the best first home if you only want to casually try one or two prompts and never use the automation side. In that case, a local Linux machine or WSL install may feel easier.

What you need before you start

Before you do anything, make sure you have:

  • a Linux machine with shell access

  • a user account that can install software

  • curl available

  • an internet connection

  • at least one model/provider you plan to use later

If you do not have a provider picked yet, that is fine. You can still install Hermes first and decide during setup.

What this tutorial will do

By the end of this guide, you will have:

  • installed Hermes Agent

  • confirmed the hermes command works

  • located the main config and environment files

  • run the first setup flow

  • sent a simple test prompt

  • checked the installation with hermes doctor

Step 1: Run the official installer

The official installer is the simplest starting point on Linux.

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

This downloads the official install script and runs it. On current builds, the installer uses a managed Python environment and sets up the Hermes CLI for you.

If everything goes well, you should finish with the hermes command available in your shell.

Note: If your shell does not pick up the command immediately, open a new terminal session or run exec $SHELL before assuming the install failed.

Step 2: Confirm Hermes is installed

Now check the installed version.

hermes --version

Hermes should return a version string similar to this:

Hermes Agent v0.16.0 (2026.6.5)

If you get command not found, the install usually completed in a location your current shell has not reloaded yet. Open a fresh shell and try again.

Step 3: Look at the main CLI entrypoint

Before jumping into setup, it helps to confirm the main command tree is available.

hermes --help

You should see a large command list that includes things like:

  • chat

  • setup

  • model

  • doctor

  • cron

  • gateway

  • config

  • tools

  • skills

That tells you the CLI is alive, not half-installed, and ready for first configuration.

Step 4: Find where Hermes stores its config

Hermes keeps its main files under your home directory.

First print the active config path:

hermes config path

Then print the active environment file path:

hermes config env-path

They should resolve to paths under your Hermes home directory, for example:

/root/.hermes/config.yaml
/root/.hermes/.env

The first file stores configuration. The second file usually stores API keys and other secrets.

Do not paste secrets into random notes or shell history if you can avoid it. Keeping them in the right place from the start saves future cleanup.

Step 5: Run the setup wizard

For most people, this is the easiest next step:

hermes setup

Unless you want to use Nous Portal by default, you might want to select Full setup here. If Hermes is already installed and you run this later, it reopens the setup flow with your current values as defaults. That is useful when you want to change provider settings, tool choices, or other basics without digging through config files manually.

If you specifically want the one-shot Nous Portal path, Hermes also supports:

hermes setup --portal

That route is designed to log you in through Nous Portal, set a Nous model, and opt into the Tool Gateway in one pass.

Step 6: Pick a provider and finish first-time setup

During setup, Hermes will guide you through the important choices. Depending on the path you choose, that may include:

  • selecting a provider

  • selecting a model

  • choosing tool settings

  • confirming gateway or agent settings later

Do not overthink every option on day one. The goal is not to build your perfect forever configuration. The goal is to get to a working first session.

If you change your mind later, you can revisit the wizard with:

hermes setup

Or jump directly into model selection with:

hermes model

Step 7: Start your first Hermes session

Once setup is done, launch Hermes interactively:

hermes

If you prefer a one-shot test instead of dropping into chat immediately, use:

hermes chat -q "Say hello and tell me which tools you can use."

That is a good quick confidence check because it verifies the CLI can actually start a session instead of just showing help text.

Step 8: Run a health check

Now do a basic sanity check:

hermes doctor

This command checks your Hermes setup and points out missing or broken pieces. It is one of the best commands to remember early because it helps with:

  • provider problems

  • dependency issues

  • general setup confusion

If Hermes suggests a fix and you understand what it is changing, you can also try:

hermes doctor --fix

I would still read the output first instead of hammering --fix on autopilot. Sometimes the boring careful approach is the one that saves you from a second problem.

What success should look like

At this point, a good result looks like this:

  • hermes --version works

  • hermes --help shows the full command tree

  • hermes setup completes without getting stuck

  • hermes chat -q "..." returns a real answer

  • hermes doctor does not report a fatal setup problem

If all of that works, you have a real baseline install, not just a nice intention and a half-working binary.

Common problems and quick fixes

hermes: command not found

Usually your shell has not reloaded after installation.

First reload your shell:

exec $SHELL

Then check Hermes again:

hermes --version

If that still fails, open a completely new terminal session.

Setup feels confusing and you want to start over

You can rerun the wizard:

hermes setup

If you only want to revisit missing or incomplete items on an existing install, Hermes also supports:

hermes setup --quick

Provider credentials are wrong or incomplete

Open the model selector:

hermes model

Correct the provider or model settings, then run:

hermes doctor

You changed something in config and it does not seem to apply

Hermes often needs a fresh session after configuration changes. Close the current session and start a new one rather than assuming the config was ignored.

How to update Hermes later

If you want to update Hermes Agent with the latest changes, run:

hermes update

On the test system used for this article, Hermes itself reported an update was available, so this is a real command worth remembering.

Where the important files live

The exact paths can vary, but the main ones are usually under ~/.hermes/:

  • config.yaml for configuration

  • .env for secrets and API keys

  • sessions/history under the Hermes home directory

  • logs under the Hermes home directory

If you are unsure which config file is active, do not guess. Run:

hermes config path

and then:

hermes config env-path

Use the printed paths rather than guessing.

If you are building this out beyond a basic install, the most useful follow-up tutorials are:

  • How to configure models and providers in Hermes Agent

  • Hermes Agent troubleshooting guide

  • Hermes Agent tools explained: what to enable and when

  • How to set up scheduled jobs in Hermes Agent

If your site already has these related foundations, this article should also cross-link to:

  • Docker on Ubuntu or Debian

  • Docker Compose for self-hosted apps

Those are not required for a basic Hermes install, but they become useful later if you decide to containerize or automate surrounding services.

You’re done

You now have Hermes Agent installed, the config paths identified, the setup wizard completed, and a basic session test in place.

That is enough to move on to the more interesting stuff:

  • choosing the right provider setup

  • enabling the right tools

  • running scheduled jobs

  • connecting Hermes to the workflows you actually care about

And honestly, that is the point where Hermes starts becoming useful instead of just theoretically cool.

Frequently Asked Questions

Does Hermes Agent only work on Linux?
No. Hermes also supports macOS and WSL. Linux is just a very practical starting point for server-style and always-on usage.
Should I use `hermes setup` or `hermes setup --portal`?
If you want the easiest Nous Portal path, use `hermes setup --portal`. If you want the more general setup flow, use `hermes setup`.
Where do I put API keys for Hermes?
Typically in the Hermes environment file. You can print its exact path with `hermes config env-path`.
What is the best first test after installation?
A quick one-shot prompt is usually enough: `hermes chat -q "Say hello and tell me which tools you can use."` Then run a broader health check with `hermes doctor`.

Related articles

🔒
Security & Firewalls
How to Install Vaultwarden with Docker Compose
Install Vaultwarden with Docker Compose on a Linux server, create your first password vault account, and lock the setup down with sane defaults for signups, admin access, and backups.
🐳
Docker & Containers
How to Install Docker on Ubuntu and Run Your First Container
Install Docker Engine on Ubuntu using Docker's official apt repository, verify that it works, and optionally set it up so you can run Docker commands without sudo.
🤖
AI & LLMs
How to Install Ollama on Ubuntu and Run Your First Local LLM
Install Ollama on Ubuntu, verify the systemd service, pull your first model, and test a local LLM from the terminal before connecting it to tools like Open WebUI or Hermes Agent.
📊
Monitoring
How to Install Uptime Kuma with Docker Compose
Install Uptime Kuma with Docker Compose on a Linux server, create your first admin account, and start monitoring websites or services from a simple web dashboard.