01What Hermes Workspace Is and How It Works

Hermes Workspace is a free, open-source dashboard that acts as a command center for your Hermes Agent. Instead of typing messages into a terminal and getting answers, you get a full desktop with memory, skills, a live dashboard, scheduled tasks, and multi-agent tools.

Think of Hermes Agent as the employee and Hermes Workspace as the desk where that employee works. The terminal is like having a full team of AI employees but only letting one of them work. Hermes Workspace lets them all work together.

In this episode, Julian Goldie shows how it turns a simple chat into a powerful system that can browse the web, write code, manage files, and remember your business context. It runs locally on your computer, and everything is fully synced with Hermes Agent.

02Main Features of Hermes Workspace

Hermes Workspace includes several sections that make managing AI agents much easier than using the terminal.

  • Chat: Talk directly with your Hermes Agent, give it tasks, and see it respond with actions, like creating videos with hyperframes.
  • Files: View and edit all files in one place, without jumping to your Mac's Finder.
  • Terminal: Run terminal commands directly inside the workspace, like installing new skills from GitHub.
  • Jobs: Set up scheduled tasks with a name, schedule, prompt, skills, and delivery options (local, Telegram, Discord). You can set a limit on runs or let it run forever.
  • Kanban: Create tasks and assign them to agents, like a project board. For example, you can say 'create an SEO calculator' and assign it to an agent.
  • Conductor: An office view where a team of AI agents works together on a mission you give them.
  • Operations: A persistent agent team that works together on ongoing work.
  • Swarm: Run unlimited agents in parallel. Julian shows 17 agents (profiles) working together.
  • Memory, Skills, MCP, Profiles: Manage what your agents know about you, add skills, and set up profiles.
  • Hermes World: A 3D game where your AI agent can interact with other AIs and even play quests.

03How to Set Up Hermes Workspace

Setting up Hermes Workspace is very simple. Julian shows that if you can copy and paste, you can do this.

  1. Copy the install command from the video description or the Hermes Workspace website.
  2. Open your terminal and paste the command. Press Enter and wait for it to install.
  3. Once installed, type 'launch Hermes Workspace' in your terminal.
  4. Open the local address in your browser (it's at localhost:3000).
  5. If you want to change the model, type '/model' in the chat and pick one, like GPT 5.4.

Julian notes that the installation takes a few minutes. He recommends using the Owl Alpha free API (O-W-L space Alpha) if you want to run it for free. You can also connect local models like Olama.

04How to Use Hermes Workspace with Examples

Once it's running, you can do a lot inside the workspace. Julian walks through several examples.

In the chat, you can say 'create a video with hyperframes, short animation by AI SEO' and the agent will actually use hyperframes to generate a video locally. You can see the video file appear in the files section.

To install hyperframes from GitHub, you can either copy the terminal command and paste it into the terminal, or give the agent the full GitHub context in the chat and ask 'set up Hermes with hyperframes.' Julian prefers the second method because it gives the agent all the information and reduces mistakes.

You can create jobs in the Jobs section. Give it a name, set a schedule, add a prompt, and choose delivery (Telegram or Discord). You can set it to run five times or every day forever.

On the Kanban board, you can add a task like 'create an SEO calculator' and assign it to an agent. The agent starts working on it immediately, and you can check back later.

In the settings, you can set up smart routing. For example, you can use Claude as the main brain and route simple queries to a local model like Olama or GPT 5.4. You can also set up voice options (Edge, Eleven Labs, OpenAI) and change the theme.

05Results and Comparisons in the Episode

Julian compares Hermes Workspace with the native Hermes dashboard. The native dashboard can't be chatted with directly, and the UI is not as nice. He shows that putting a task on the Kanban didn't work properly there.

He also compares it to Aion UI. Aion UI is good for managing multiple agents, but it's like a ChatGPT for agents. Hermes Workspace has more functionality built specifically for Hermes, like Swarms and Kanban.

So for Hermes users, he recommends Hermes Workspace.

Julian shows a real result: a 450-page website created by an agentic swarm inside Hermes Workspace. The website is fully internally linked, has FAQs, tables, and looks better than most humans can do. It runs locally and was coded locally using agent swarms.

This proves that with the right systems, AI can do real work, not just generic answers.

06Tips, Limits, and Mistakes to Avoid

One mistake is giving the agent only a small command when installing from GitHub. Julian recommends giving the full GitHub context to avoid errors.

Another limit is the native Hermes dashboard—it can't chat or manage agents well. Use Hermes Workspace instead.

Some people think they're not technical enough. Julian says if you can copy and paste, you can do it. He's seen 60-year-olds set it up easily.

It's free forever, so there's no cost barrier. You can even use free APIs like Owl Alpha or local models.

One thing to remember: smart routing is powerful, but you should confirm with Hermes how to use it. Tell it in the chat what model to use for what tasks, so it understands your setup.

Don't just use AI like a calculator. That's like buying a Ferrari and only driving to the corner shop. Set up the system to do the work for you.

07Who Hermes Workspace Is For and Why It Matters

Hermes Workspace is for anyone who uses Hermes Agent and finds the terminal too limiting or ugly. It's perfect for entrepreneurs, SEOs, and content creators who want to scale their work without doing everything manually.

Why it matters now? We're in the biggest technological shift since the internet. Early movers will have the biggest advantage.

This tool let you set up a system where AI agents work together, remember your business, and take real actions.

Julian's Goldie Agent Empire framework has five layers: command center (Hermes Workspace), brain (using Obsidian and OMI for context), skills stack (pick 5-10 core skills), automation engine (use swarms to stop being the operator and become the architect), and swarm (multiple agents research, write, edit, publish).

It can save you 10 hours per week. That's 60 hours a month. If you spend one hour learning it, it's worth it.