Install Hermes and Run It for the First Time

This is the hands-on lesson. You will install Hermes Agent with a single command, connect it to a language model so it can think, and start it for the first time. The steps are beginner-friendly. You will paste one command and answer a few prompts. Take it slowly and read each step before you run it.

What You'll Learn

• What you need before you install (and what you do not)
• The single command that installs Hermes
• How to connect a language model backend so Hermes can think
• How to start Hermes and say hello for the first time
• Where your data is stored on your machine

Before You Start: What You Need

Hermes runs on Linux, macOS, and Windows through WSL2 (Windows Subsystem for Linux). You do not need to install programming languages or databases yourself. The installer sets up what it needs automatically.

You will need three things:

1. A supported machine. A Mac, a Linux computer, or Windows with WSL2 enabled. If you are on Windows and do not have WSL2, search for "install WSL2" and follow Microsoft's official guide first. WSL2 gives you a Linux environment inside Windows.
2. A terminal. This is the text window where you type commands. On macOS it is the Terminal app. On Linux it is your shell. On Windows it is the WSL2 window.
3. Access to a language model. Hermes orchestrates, but the actual thinking comes from a large language model. You will connect one during setup. We cover the options below.

That is it. No coding background required.

Step 1: Run the Install Command

Open your terminal and paste the official install command from Nous Research:

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

Press Enter. This downloads the installer and runs it. It will set up Hermes and the things it depends on for you. This can take a few minutes the first time. It is normal to see a lot of text scroll past.

A quick word on safety, since you are piping a script from the internet into your shell. This is the install method the project documents, and it is common for developer tools, but it is also worth understanding. If you want to be cautious, you can open the URL in a browser first and read the script before running it. That is a reasonable habit for any "curl into bash" command, not just this one.

Step 2: Connect a Language Model Backend

After installing, Hermes needs a brain. It supports several backends, so you can pick what fits your budget and privacy needs:

• Nous Portal. The first-party option from Nous Research. You sign in with an OAuth login (the same kind of "log in with..." flow you have used on other sites). This is often the simplest starting point.
• OpenRouter. A single service that gives you access to 200+ models from many providers through one API key. Handy if you want to try different models without separate accounts.
• A custom OpenAI-compatible endpoint. If you already use a provider that follows the common API format, you can point Hermes at it.
• Local vLLM. This runs a model on your own hardware. It keeps the thinking on your machine too, which is the most private option, but it needs a capable computer (a strong GPU helps).

For most beginners, Nous Portal or OpenRouter is the easiest way to start. Whichever you choose, follow Hermes's setup prompts and paste in the key or complete the login when asked.

A note on cost and privacy. If you use a cloud backend (Nous Portal, OpenRouter, or a custom hosted endpoint), your prompts are sent to that provider, and you pay it for usage, usually a small amount per action. If keeping everything on your machine matters to you, the local vLLM option does that, at the cost of needing stronger hardware. There is no single right answer. Pick based on your budget and how sensitive your data is.

Step 3: Start Hermes and Say Hello

Hermes gives you two ways to interact with it.

The quickest is the built-in terminal interface. Start it by typing:

hermes

This opens a chat right in your terminal. Try a simple first message, such as asking it to introduce itself or to summarize a short piece of text you paste in. This confirms everything is connected.

The second way is the gateway, which lets you talk to Hermes from a messaging app. Once you start the gateway and connect a platform, you can message Hermes from Telegram, Discord, Slack, WhatsApp, or Signal, just like messaging a contact. The terminal interface is the fastest way to confirm your install works, so start there. Connect a messaging platform later once you are comfortable.

Step 4: Know Where Your Data Lives

Everything Hermes stores, its memory, the skills it writes, and its configuration, lives in a single folder in your home directory:

~/.hermes/

The ~ is shorthand for your home directory, and the leading dot means the folder is hidden by default. You can list its contents to see what is there. Because it is all local and readable, you have full visibility into what your assistant is keeping. This is also the folder to back up if you want to preserve everything Hermes has learned, and the folder to inspect if you ever want to review or clean up its memory.

If Something Goes Wrong

A few common first-run issues:

• The command says "command not found" when you type hermes. The install may not have finished, or your terminal needs to be reopened so it picks up the new command. Close and reopen the terminal, then try again.
• Hermes starts but cannot think. This usually means the language model backend is not connected. Re-check the key or login you set up in Step 2.
• You are on Windows and the command fails. Make sure you are running it inside a WSL2 terminal, not the regular Windows Command Prompt or PowerShell.

When in doubt, the official documentation at the Hermes Agent site is the authoritative source, since installer details can change between versions.

Key Takeaways

• Hermes installs on Linux, macOS, and Windows (via WSL2) with a single curl command. No coding background needed.
• After installing, you connect a language-model backend: Nous Portal, OpenRouter, a custom endpoint, or local vLLM.
• Cloud backends cost a small amount per action and send prompts to the provider; local vLLM keeps thinking on your machine but needs stronger hardware.
• Start with the terminal interface by typing hermes, then connect a messaging platform later.
• All of Hermes's data lives locally in ~/.hermes/, which you can inspect and back up.