01What Hermes Agent Is and How It Works

Hermes agent is an open-source AI tool that turns a basic chat interface into a fully autonomous system. It can run tasks, build projects, and work in the background without you managing every step. In this episode, Julian Goldie explains how to unlock more power from Hermes agent.

Hermes agent works by following instructions you give it, either directly in the terminal or through files like the sole MD file. The sole MD file acts like a custom prompt that controls how the agent thinks and responds. By editing this file, you can make the agent more proactive and autonomous.

Julian shows that even small changes in this file can completely change how the agent behaves.

You can run Hermes agent locally on Mac or Windows, or on a cloud host like Hostinger for better performance. This flexibility makes it useful for beginners and advanced users alike.

02Main Features and What’s New

Hermes agent offers several powerful features that go beyond simple chat. The most important ones shown in this episode include:

  • The sole MD file, which lets you customize the agent’s thinking and response style.
  • Forward slash goal, a command that creates an autonomous persistent loop for one-off tasks, using up to 20 loops to complete them.
  • Hermes Workspace, a free extension with a user-friendly interface for managing files, memory, skills, and more.
  • Agent swarms, which let you run multiple agents in parallel to work on big projects.
  • Integration with local LLMs like LLM Studio and Ollama for free, without API keys.

In this episode, Julian also mentions that Hermes agent has a growing open-source community, with many new projects like Aion UI and multicore coming out regularly. This makes it a vibrant ecosystem to be part of.

The key new feature is the ability to manage everything through Hermes Workspace instead of the terminal. This makes the tool much more accessible to non-developers.

03How to Set Up Hermes Agent

Setting up Hermes agent is straightforward. Julian shows the steps he uses, both for local setup and for better performance on cloud hosting.

  1. Go to the Hermes agent GitHub page and follow the quick start instructions.
  2. If you are on Windows or Mac, copy the install command and paste it into your terminal.
  3. For local LLMs, run the setup command (like 'hermes setup') in your terminal and choose LLM Studio or Ollama.
  4. If you want a nicer interface, install Hermes Workspace from GitHub.
  5. For better performance on a Mac mini, use a cloud host like Hostinger instead of running locally.
  6. Once installed, ask the agent 'where is my sole MD file?' to find and edit it.

Julian recommends not to restart or delete your setup if it feels slow. Instead, improve it by editing the sole MD file or asking the agent directly how to make it more autonomous.

04How to Use Hermes Agent for More Power

To get the most out of Hermes agent, you need to customize it. Julian shows several ways to do this.

First, edit the sole MD file to include instructions like 'be genuinely helpful, have opinions, be resourceful, be autonomous, be more proactive.' This changes how the agent thinks and responds. You can also ask the agent directly, 'how can I make you more autonomous?' and then apply the suggestions.

For one-off tasks, use the forward slash goal command. This creates an autonomous loop that keeps working until the task is done, with up to 20 loops available. This is great for complex tasks that need iteration.

To use a local LLM, type forward slash model in the chat and select your model. Make sure LLM Studio or Ollama is running in the background. You don't need an API key; just use the base URL and provider name.

To run multiple agents in parallel, open Hermes Workspace and go to the Swarms section. You can choose from pre-made agents or create your own. This lets you build big projects, like a 50-page website, quickly.

05Results and Comparisons from the Episode

Julian shares real results from using Hermes agent. He mentions that he and his team built a full 50-page website using agent swarms. This shows how powerful parallel agents can be for large tasks.

He also compares Hermes agent to OpenCore. Julian says he switched to Hermes because it is easier to use, works more reliably, and has a better open-source community. He notes that OpenCore breaks often with updates, while Hermes works smoothly.

Since switching, he says he uses AI agents more often and gets more powerful use cases out of them.

Another result is from a community member named Joe, who deleted OpenCore, installed Hermes agent and Hermes Workspace, and called it a 'huge win for the weekend.' Joe said he finally feels like he is accomplishing something.

06Tips, Limits, and Mistakes to Avoid

There are some common mistakes to avoid when using Hermes agent. Julian shares these tips from the episode:

  • Do not restart or delete your setup if it feels unresponsive. Instead, improve the sole MD file or ask the agent for suggestions.
  • Use cloud hosting like Hostinger if you have performance issues on a Mac mini.
  • For Windows setup problems, use Claude Code to help you follow the GitHub instructions step by step.
  • Make sure to configure LLM Studio or Ollama in the setup before trying to use local models.
  • Keep LLM Studio running in the background when using local LLMs.
  • Use agent swarms for big projects, but avoid using them for simple daily tasks where forward slash goal is better.

One limit is that Hermes agent is not as proactive as you might expect by default. You need to edit the sole MD file to make it more autonomous. Also, the terminal interface can be intimidating, but Hermes Workspace solves that.

07Who Hermes Agent Is For and Why It Matters

Hermes agent is for anyone who wants to use AI to automate tasks and run projects without constant supervision. This includes non-developers like Karen, a health coach who wants to build a personal AI system for her business.

Julian explains that Hermes agent is perfect for people who want to create an 'agent team' that runs in the background. You can set up different profiles for different purposes, like marketing or support, and even connect them to different Telegram groups.

Why does this matter? Because it lets you focus on other work while AI handles repetitive tasks. By using Hermes Workspace and agent swarms, you can build websites, manage projects, and answer questions automatically.

This makes it a valuable tool for businesses and individuals looking to scale their productivity.