NEW Hermes Jarvis is INSANE! artwork

NEW Hermes Jarvis is INSANE!

AI News Today | Julian Goldie Podcast

June 6, 2026

Hermes Jarvis: Voice-Control Your AI Agent to Run Your Computer & Automate WorkThe script introduces “Hermes Jarvis,” a voice interface that connects a microphone to the Hermes AI agent so spoken commands can trigger real actions on a computer, not just conversation.
Speakers: Julian Goldie
**Julian Goldie** (0:00)
Hermes Jarvis is absolutely insane. So this is a way that I can control my AI agent with my voice to, for example, control my laptop or to control browser use. Let me show you an example right here. So if we just speak to it like so, hey, can you type in a new note inside my notes? Just give me a list of automations I could create. So you can see it's created a new note now, and now it's actually written the notes inside the project as well. So we've got all the notes created inside the app.

**SPEAKER_2** (0:27)
I've created a new note in your notes app with a comprehensive list of automation ideas across nine categories, content, social, video, audio, SEO, business, AI, dev, productivity, smart home, plus tools to build with. The note is saved and ready to use.

**Julian Goldie** (0:45)
So you get the point. You can speak to your AI agent now. It can control your computer. You can get it to do whatever you want, basically fully automated, and you don't even need to type to it anymore. You can just voice activate it, and from there it can actually go off and do things. And this is what I call Hermes Jarvis. And this basically connects your voice to your AI agent, and then what it will do afterwards is read it out loud. And we built this into our agent operating system. So you've probably seen voice AI systems to just chat about, but this is different because Hermes Jarvis is actually wired a mic to the real Hermes agent, the one that can actually do stuff. So you speak a job out loud, for example, I don't know, research my top competitors, write me a brief, be on the aligning page, find the top three AI news stories, it goes and does the actual work, then it talks the results back to you. And today I'm going to show you how it works. Let me show you another example here. Schedule researching the latest AI automation news daily at 3pm.

**SPEAKER_2** (1:43)
I've created a daily cron job that runs at 3pm to research the latest AI automation news. The script saves results to AI automation news, daily, research, log and individual json files by date.

**Julian Goldie** (1:57)
So you can see how it works. You can basically schedule tasks. You can get it to use a computer. You could get it to use bronzies, whatever you want using this AI agent. Super powerful. Also just a lot of fun to play with. The difference here is that normally a chat would just act. If you were just like, for example, going back and forth with chat chibity inside the voice assistant, it can't really go off and do things for you.
Jarvis with Hermes agent is really the moment where agents and voice come together and you speak and the work gets done. Now, most AI voice only talks. Here's the gap nobody talks about.
Voice assistants you know can answer questions, but they can't actually do anything. You can ask one to be on the page or go research this, etc. It just describes how you might do it. You still have to do the work. This is an AI that can actually go off and do stuff for you. And it's an assistant wired to the agent with hands, right? So we've basically got this tab here with multiple different voice options. You can see we can switch between these agents. This is actually using the voice commands from Minimax M3. And then these are plugged straight into our actual Hermes agent, the one with all the tools, which means whatever Hermes can do, you can now trigger with your voice. So that could be like searching the web, reading and writing files, building work in pages and apps, running its skills. And so I call this Goldie Jarvis mode, which is you say it, you do it, you hear it. So the whole thing is three simple moves.
You can tap the mic and speak what job you want done in plain English, just like talking to a capable assistant.
No special wording, you don't need to type or anything. It's going to go to work and start researching or building or writing files, even using a browser if you want to. So it's a natural doer, not a describer. And then hear it. So once it's done, it actually talks the result back to you in an actual voice. So you can keep your eyes and hands on something else, right? And so you could ask something out loud, go take a walk, come back to the finished file, right? For example, we could say, build out a to-do list app. There's just simple HTML and then open it up for me. And so that will just go out. You can see here that we've given it the voice command and then it will just go to work in the background. Now, the honest part here is that you just got to use it wisely, because, you know, it can do real things. You want to treat it like a keen new assistant, literally takes you literally, OK? So it acts on what you say, screen control and that sort of thing. It's opt in here and also is not going to be like instant. So you can see, for example, here, we've said go out and build a to do list app. It might take a couple of minutes to just code that out for us, but it's working in the background whilst we go off and do something else, which is pretty cool as well. So if you want Hermes set up and ready to go, we've got it inside the agent OS. Otherwise, you can actually just build this with Hermes directly. You just have to make sure you have the right API set up and everything else. We've set up, for example, Minimax M3 with Hermes agent, and that's how this whole thing is running together. Then on the actual model of Hermes, we've actually used something called Nemotron that just came out today. So some people want to say voice AI is a gimmick, it just answers trivia, but this one is actually wired to real agents. It can research, it can build, it can save files from a spoken sentence, and it actually does work. Other people say, well, talking to AI is slower than doing it myself. The thing is, you can speak a sentence and then walk away, right? So you can do all the multi-step work, all the research, all the scheduled tasks in the background whilst you're doing something else. Other people say, well, an AI that controls my computer sounds scary. So for example, like computer use or anything like that is opt-in. So usually when you're using Hermes agents to do stuff like this, you actually ask you for permissions first before you approve it, and that's basically it. So just to recap there, Hermes, Jarvis, most voice AI only talks. Hermes, Jarvis actually wires your voice to the real Hermes agent. Pretty powerful stuff. Spoken command can just go off to do things. So this is working on that in the background. By the way, there's so much more you can do with Hermes as well. So for example, we've actually got a live chat with it as well, where we can actually just have a live conversation and go back and forth. We have the Jarvis section here. We also have the studio where we can generate images and videos and voice, and we have the sessions where we can manage everything and see the status. So you can see here, we've got Nvidia Nemotron 3 Ultra plugged in. We've got the Gold Mode section set up. We've got all of this different stuff that we've created, and we can just come back to our workspace, see what we've built, etc.

2 more minutes of transcript below

Feed this to your agent

Try it now — copy, paste, done:

curl -H "x-api-key: pt_demo" \
  https://spoken.md/transcripts/1000651996090

Works with Claude, ChatGPT, Cursor, and any agent that makes HTTP calls.

From $0.10 per transcript. No subscription. Credits never expire.

Using your own key:

curl -H "x-api-key: YOUR_KEY" \
  https://spoken.md/transcripts/1000771488799