Agent OS: Local Setups, VPS & Higgsfield! artwork

Agent OS: Local Setups, VPS & Higgsfield!

AI News Today | Julian Goldie Podcast

July 18, 2026

Agent OS: Local Setups, VPS & Higgsfield!
Speakers: Julian Goldie
**Julian Goldie** (0:00)
Today, we're gonna be answering some of the latest questions about our agent operating system, how to build, how to get the most out of it, and how to basically automate anything. Now, what I do every single day inside our AI Profit Boarding community is we basically look at the latest questions, the latest challenges that people are facing, and then we solve them on a daily video tutorial like this. So, we're gonna get straight into this, and we've got the latest questions inside the community over here, so let's take a look. So, the first one is from Kevin, and he's actually redone his whole sales funnel for his agency using AI, which is pretty impressive in itself in just a week, but he's also talking about, okay, I need to figure out how to take images and turn them into like amazing videos with the least amount of input possible.
So, there's a few options for this. One of the best options for this is you can actually get the API from HeyGen, and HeyGen have an image to video AI generator, so that's one good option. Another option is that you can actually use Higgs Field. Higgs Field is a great tool for generating beautiful videos from an image, and they also have an MCP. So, if you want very little input into creating the video, so you just want your agents to handle them for you, you can get the MCP like so. It's only like three steps to set it up, and then you can plug that into any agent that can use MCP. So, that could be Hermes, could be Codex, could be ChatGBT, could be, for example, Claude, whatever you prefer. The main point here is like, you'd create a skill with your AI agent that understands your brand, and then you connect your API or MCP to the agent so that it can take the image and generate a video using your skills so that it understands how to brand the videos correctly too. We've got a question here about how to basically take your agent OS, which you can see here as an example, how to take the agent OS and then plug it from a VPS over to Windows. So right now, he's running everything on a Linux VPS.
They'd like to move the entire project to Windows and continue developing it locally with Claude Code. So I actually personally don't use Windows, but there's some good stuff here. So Michael actually said one option is GitHub if you aren't already using it. So you can basically keep your VPS and use a local environment synchronized because you can basically commit the whole project to GitHub, and then they can be shared across the local and the VPS environment, even if you're using Windows. I think that's probably the best solution that I can see here. And it's fantastic that we've got so many helpful people inside this community just helping each other out and trying to solve problems together. We've got a question about running the Agent Operating System and Paperclip. Now, if you're not familiar with Paperclip, basically this is an open source team orchestration system. So you can orchestrate all of your agents using Paperclip. Now, the question here from Scott is Agent OS and Paperclip. Any thoughts on running both? So I actually have Paperclip built into the Agent OS, and that's available for everyone who sets it up from us.
Zipfile inside the AI Profile Boardroom. I honestly don't use it that much, just because I tend to find the AI agents already very good at orchestrating themselves anyway, and that you typically don't need a full team. However, I do think it's useful for visualizing how a team of AI agents can work together. So if you want like something really visual, where you can see your agents working together in a team, you can assign them different responsibilities, etc. Then I do think Paperclip is useful, but just in reality, I tend to find like using the workflows that we have individually, like Hermes Apollo, Hermes Oracle, Hermes Asteros, etc.
They tend to be way more useful and practical than having something like Paperclip, which can also be a bit buggy sometimes, honestly, as well. Now, something you'll notice about our setup with the AgentOS is that it's hosted on a local host, like so. One of the questions we got here from Eric was how to basically restart it. So basically, if you shut down your computer, sometimes when you reload it, the local host URL is down. And if that happens to you, what you can do is just ask the agent that set up your AgentOS to restore and restart the local host. And sometimes I just say, hey, AgentOS isn't open. Can you just load it back up for me? And then it gets out pretty quickly again. Otherwise, what can happen is like sometimes you have a local port, but it's not being used for the AgentOS after you restart your computer. So that's pretty much it when it comes to the latest questions or answered every single question inside the community today. If you need any help like this or you want to ask me questions and I create a video tutorial for you like this, then you can post your questions inside the community section here. Just write your post and I will answer it for you with a video tutorial.

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/1000777283807