Hermes GPT 6 is the setup I talk to every morning, and it has built me more working software in the last few months than I managed in years of hiring people to do it.
I don't code.
That's not false modesty, it's the whole point of this article.
Below are nine things this setup actually does for me, in the order I use them, with the honest limits attached.
Here's the system running, then we'll go through what it produces.
1. It builds interactive apps from a sentence
The first thing I ever asked for was a galaxy I could swirl with my mouse.
It said "built and running, sir — have a look".
And there it was, working, in the browser.
That's the use case that reframes everything else, because once you've watched a spoken sentence become a running app, you stop thinking in terms of "can it" and start thinking in terms of "what next".
2. It builds tools I'd never have paid a developer for
There's a habit tracker in my gallery that I use.
There's a meditation timer.
There's a Japanese flashcard game, which exists purely because I wondered aloud whether it could.
None of those were worth a developer invoice or a month of my time.
All of them were worth ten minutes of talking, and that's the economics that changed.
3. It builds full websites
One of the entries in my build list is a complete website.
Not a wireframe, not a mockup, a site.
And let's be honest about the trajectory here — this is the worst it is ever going to be.
4. It creates videos
It can even create videos, which still catches people off guard when I show it.
Combine that with the build gallery and you have a place where every asset you've generated is sitting waiting for you.
Here's the wider Agent OS build where a lot of this comes together.
🔥 Want the exact setup I used to get these results? Inside the AI Profit Boardroom, I've got a full Hermes agent section with step-by-step video tutorials plus the Agent OS install zip and prompts. Weekly coaching calls + 2,800+ members building real automations. → Get access here
5. It tells me what I've been working on
I ask "Jarvis, what have I been working on this week?".
It tells me I've been tinkering with thumbnail work, because the thumbnails log shows recent edits there.
That's not a party trick, that's a weekly review I didn't have to write.
It works because my Hermes conversations are linked to my Obsidian memory, which I call the memory galaxy.
Everything I do with Jarvis lands in a log I can come back to later, which just makes managing all of it easier in one place.
Here's that second brain layer on its own.
6. It pulls up my numbers on request
"Jarvis, show me my numbers."
"Pulling up the numbers, sir."
Small use case, used constantly, and exactly the kind of thing that dies in a normal chat window because the model can't reach anything.
7. It opens things for me
I say "Jarvis, open up juliangoldie.com" and it opens the site.
Computer use and browser use are built into the system, which is powerful in itself.
The version of this I care about most is verification — an agent that can open a page can check the thing it just built actually loads.
8. It shows me my agent team
"Jarvis, show me my team."
It pulls up the agents ready inside the system — Claude, the open core agents, Hermes, all together.
From there I've got mission control, goal mode for outcome-driven work, chat mode for typing, and MCPs for adding anything that isn't built in.
I covered that control layer in Hermes agent mission control.
9. It lets my models argue with each other
Paperclip is built in with AI agent teams.
There's also an agent group chat where Hermes can talk to Claude and Gemini.
That particular part is my favourite bit of the entire build.
Different models pressure-testing each other beats any single model's first answer, and I wrote about the multi-agent side in Hermes agent swarm.
What the stack is underneath
There are three parts to this, and none of them are mysterious.
Hermes agent is the open source project, and all of its agentic capabilities are built into the system.
ElevenLabs generates the voice response.
Claude works in the background on the improvements and the system itself.
The model you select in the model manager is the brain, and that's the slot the "GPT 6" part of the name refers to.
Swap that slot and everything in this article gets better without you touching anything else.
| Use case | What you say | What you get |
|---|---|---|
| Build an app | "Build me a galaxy I can swirl" | A running page |
| Weekly review | "What have I been working on?" | An answer read from your logs |
| Open a site | "Open up juliangoldie.com" | The browser opens it |
| See past work | "Show me everything I've built" | The full build gallery |
| See your agents | "Show me my team" | Claude, open core, Hermes |
| Check metrics | "Show me my numbers" | Your numbers, spoken |
Where it falls short
It doesn't always work.
Sometimes it's not that smooth, and I'd rather you hear that from me.
If your goal is pure implementation volume, I'd actually point you at paperclip or an agent teams setup instead, because you create a lot more a lot faster that way.
The voice Jarvis build is my favourite party trick with AI and it's genuinely useful, but those are two different jobs and you should pick per job.
Also worth knowing: I run a parallel version on free Claude Code with an agent factory, using Qwen 2.5 Coder through Ollama as a free local model.
That runs entirely on my own machine, which tells you how far local models have come.
🔥 Want my full Hermes Jarvis + Agent OS setup? The AI Profit Boardroom includes the full video tutorial, the install zip with prompts, daily new tutorials, four weekly coaching calls where you share your screen and get help live, and 2,800+ members. → Get the setup here
Hermes GPT 6 FAQ
What can Hermes GPT 6 build?
Working apps and sites.
My own gallery has an interactive galaxy, a habit tracker, a full website, a Japanese flashcard game and a meditation timer, plus generated videos.
Can Hermes GPT 6 control my browser?
Yes, computer use and browser use are built in.
Ask it to open a site and it opens the site.
How does Hermes GPT 6 remember things?
Through an Obsidian vault linked as a second brain, plus the full conversation history and build gallery inside the interface.
Is Hermes GPT 6 better than an agent team for getting work done?
For raw output, no — agent teams ship more, faster.
For hands-free building and for the demo that makes people sit up, the voice build wins.
Do I need to code to use Hermes GPT 6?
No.
Every build in this article came from me talking to it, and I don't code.
Can Hermes GPT 6 work with other AI models at the same time?
Yes.
The agent group chat lets Hermes talk to Claude and Gemini, and the team view shows everything available in the system.
About Julian
I'm Julian Goldie — AI entrepreneur, SEO expert, and founder of the AI Profit Boardroom (2,800+ members). I help business owners scale with AI agents, automation, and SEO.
- 400K+ YouTube subscribers
- 7-figure AI agency (Goldie Agency)
- Daily training inside the Boardroom
- Author of multiple AI automation playbooks
→ Get my best AI training inside the AI Profit Boardroom
Also On Our Network
- 🌐 Read on bestaiagentcommunity.com
- 🌐 Read on aiprofitboardroom.com
- 🌐 Read on juliangoldieaiautomation.com
- 🌐 Read on aimoneylabjuliangoldie.com
Related reading
- Hermes agent use cases — the seven I use every day.
- Hermes agent mission control — the control layer.
- Claude + Hermes agent — the improvement loop explained.
Nine use cases in, the thing I'd tell anyone weighing up Hermes GPT 6 is that the harness matters more than the model — get that right and the builds start showing up on their own.
📺 Video notes + links to the tools 👉
