Get Ready for the Hermes Workshop

Before you start

Welcome, and thank you for joining us! We're really glad you're here.

By the end of the workshop, you'll have your own Hermes AI agent running on your own machine. One that knows your context, remembers your projects, and carries on working while you sleep.

The steps below are just getting your workbench ready, so that on the day we can start building right away. It takes 10-20 minutes, and none of it is difficult.

Would rather just watch the workshop and build on your own time? Completely fine! Everything is recorded, so you can build along later at your own pace.

Steps

Step 1 · 1 min

Pick where Hermes will live

Hermes runs on a machine you control, so the first decision is simply which machine.

Our advice: just use your laptop. It's the fastest way to get going, and moving Hermes somewhere else later is straightforward; we'll send resources on how to do that after the workshop. Don't overthink this one.

Step 2 · 5 mins

Install Hermes and connect a model

This is the one step that really matters. Two parts: install the app, then give it a brain.

2a. Install Hermes

Download the desktop version from hermes-agent.nousresearch.com/desktop. It runs on macOS, Windows, and Linux, and it's free.

Prefer the command line? You can install it that way too; see the official install guide.

2b. Pick a model

Hermes is the agent. The model is the brain it thinks with, and you choose which one. Pick based on what you already pay for, or what you feel like spending.

Head's up: if you have a paid Anthropic or OpenAI subscription, make sure you're using your subscription and not the metered API!

2c. Send your first message

Pick a model in Hermes and say hello. If it replies, you're done.

You're finished with this step when Hermes is installed on your chosen machine, and a model answers you.

Step 3

Keep a chatbot open

Have a general AI chatbot open in a tab during the workshop. It'll come in handy! 😊

Already pay for one? ChatGPT, Claude, Deepseek, Grok, Gemini etc. all work fine. Just choose your go-to AI chatbot. Don't pay for one? The free tier of any of them will be good enough.

Step 4 · 5 mins

Install a coding agent

Don't let the name put you off. For most people, this takes less than 5 minutes.

A coding agent is just an app that helps you write code. You type what you want in English, and it does the rest. Don't worry if you don't know how to code. For this workshop you don't need to understand the code, and you won't be directly writing any code yourself.

If you already pay for Claude or ChatGPT

You're almost done already. Download the desktop app, click sign in, and that's it.

Claude subscription → download Claude for Desktop

ChatGPT subscription → download ChatGPT for Desktop

The videos below show you exactly what to click. Each one is about a minute.

If you'd rather not pay for anything

Use OpenCode instead. It's completely free, and it works with free models as well as the local models in Step 5.

We've made a walkthrough you can follow along with so you can set up in minutes:

One last thing

Coding agents can be accessed via desktop app, command line, through Cursor, VS Code, or any other IDE too. If you're already familiar with coding agents feel free to use whatever method you prefer. But if you've never used a coding agent before please download the desktop app version because that's what we'll be using during the workshop.

You're finished with this step when it's installed, you're signed in, and you've asked it one question and got an answer back.

Step 5 · Optional · 10 mins

Run models on your own machine

Skip this one unless you want to take it a step further and run models privately on your own computer.

Each tool below runs a small server on your machine. Hermes connects to it just like it would a paid provider; you just paste in a local address instead of signing in.

Ollama

Downloading a model with Ollama and connecting it to Hermes

1. Download Ollama from ollama.com
2. Download a model
3. In Hermes, choose a custom / local endpoint and paste http://localhost:11434/v1

LM Studio

Starting the LM Studio local server and connecting it to Hermes

1. Download LM Studio from lmstudio.ai
2. Search for and download a model inside the app

1. Open the Developer tab and start the local server
2. In Hermes, choose a custom / local endpoint and paste http://localhost:1234/v1

llama.cpp (for those who code)

llama.cpp gives you the most control and usually the best speed, but it runs from the terminal. Follow the official llama.cpp guide, then start the server:

1. Install llama.cpp using the instructions in the guide for your system (for example brew install llama.cpp on macOS)
2. Download a model in GGUF format, for example from Hugging Face
3. Start the server: llama-server -m path/to/model.gguf --port 8080
4. In Hermes, choose a custom / local endpoint and paste http://localhost:8080/v1

FAQ

Frequently Asked Questions