Hermes Agent
Apr 17, 20262

How to Install Hermes Agent

A practical install guide for Hermes Agent: the official one-line installer, first-run checks, `hermes setup`, and the few commands worth verifying before you do anything else.

How to Install Hermes Agent

If you want Hermes Agent running quickly, there is no reason to overcomplicate the install.

The official quickstart path is straightforward:

  1. run the one-line installer
  2. reload your shell
  3. verify hermes
  4. run hermes setup
  5. run hermes doctor

That is the whole core flow. This article sticks to that path and cuts out the usual filler.

Hermes Agent official quickstart page

Source: Hermes Agent quickstart

Supported Environment

Current official docs support:

  • Linux
  • macOS
  • WSL2 on Windows
  • Android via Termux in the quickstart path

The simpler rule is:

  • macOS and Linux are the easiest
  • Windows users should go through WSL2

If you are trying to treat native Windows as the default path, you are already off the official track.

The Install Command

The official one-line install command is:

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

That is still the canonical baseline in the current docs.

What the Installer Actually Solves

The installer matters because it removes the annoying parts:

  • fetching the Hermes package
  • handling runtime dependencies
  • wiring the CLI into your shell

For most users, there is no upside in manually recreating what the script already does unless you are debugging a broken environment.

Step 1: Run the Installer

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

Wait for it to finish. If you are on Linux or macOS, that is usually the entire installation step.

Step 2: Reload Your Shell

The official docs explicitly tell you to reload your shell after install.

Examples:

source ~/.bashrc

or:

source ~/.zshrc

If you skip this, hermes may be installed correctly but still appear missing in the current terminal session.

Step 3: Verify the CLI

Run:

hermes --version

If that prints a version number, the install worked.

If hermes is still not found:

  • reopen your terminal
  • reload the right shell config file
  • confirm the installer actually completed

Do not jump into provider setup before the CLI itself is callable.

Step 4: Run the Setup Wizard

Once the CLI works, run:

hermes setup

This is the step that actually makes Hermes usable. The current docs describe hermes setup as the main path for:

  • choosing a model provider
  • entering API credentials when needed
  • optionally configuring messaging
  • initializing the first working configuration

If you stop after installation and never run setup, Hermes is installed but not really ready.

Step 5: Run Diagnostics

After setup, run:

hermes doctor

This is the best sanity check before you go further into models, gateway setup, skills, or messaging.

It is especially worth doing if:

  • the install felt inconsistent
  • your shell cannot find part of the runtime
  • provider setup partially worked
  • you are on WSL2 or a less standard environment

The Minimum First-Run Command Set

You do not need to memorize a huge Hermes CLI on day one. These are enough:

hermes --version
hermes setup
hermes doctor
hermes model
hermes gateway
hermes update

What they are for:

  • hermes --version: verify install
  • hermes setup: initialize first working config
  • hermes doctor: validate runtime health
  • hermes model: change or inspect your model choice later
  • hermes gateway: configure messaging only when needed
  • hermes update: update the install

Do Not Mix Base Install with Gateway Complexity Too Early

One common mistake is trying to configure Telegram, Discord, Slack, or other channels before the base install is healthy.

The cleaner order is:

  1. install Hermes
  2. verify hermes
  3. run hermes setup
  4. run hermes doctor
  5. only then add gateway or messaging

That order makes debugging much easier.

What to Do Right After Install

If the install completed and doctor passes, your next move depends on your goal:

  • if you want a normal cloud-backed setup, keep going with provider setup
  • if you want a local model, move to the Ollama article next
  • if you want Hermes in messaging apps, configure the gateway after the base setup is stable

That is enough for a real first session.

The 5 Most Common Install Mistakes

1. Not reloading the shell

This is the classic one. The install finishes, but the current shell still cannot see hermes.

2. Treating install as the same thing as setup

Install gets the CLI on your machine. hermes setup makes it usable.

3. Trying to configure messaging too early

Get the base CLI working first.

4. Skipping hermes doctor

This removes the easiest troubleshooting checkpoint.

5. Using an unsupported Windows path

If you are on Windows, follow the WSL2 route.

Quick Checklist

  • Run the official installer
  • Reload your shell
  • Verify with hermes --version
  • Run hermes setup
  • Run hermes doctor
  • Add gateway or messaging only after the base install works

Official References