Agent OS: Obsidian + Multi-Agent Orchestration + GLM 5.2! artwork

Agent OS: Obsidian + Multi-Agent Orchestration + GLM 5.2!

AI News Today | Julian Goldie Podcast

July 1, 2026

Agent OS Q&A: Memory Galaxy Setup, Multi-Agent Orchestration, Video Agents, GLM 5.
Speakers: Julian Goldie
**Julian Goldie** (0:00)
So today, we're going to be answering the latest questions on our agent operating system. This is basically how we combine all of our agents together. So for example, we've got like Hermes and the Hermes Oracle. We've got a lead generation machine over here. We have Claude ready to go. We've got a mission control system and everything we need to win with AI inside one single place. We even got, for example, Kanban, NotebookLM, all synced inside one place. So today, what we're going to be doing is answering some of the latest questions that I've got inside the AI Profit Boardroom community, because I know if people have questions inside here, you probably have similar questions about agent operating systems, how to build them, etc. So let's get straight into this and answer some of the last questions. So a good post from Andrew and he was saying, how he's basically set up the agent OS. He was saying, huge thank you to Julian, everyone here.
Basically, what he's done is use our agent operating system and then build this himself, but he's also shared everything he's learned. So some good tips here, number one, like keeping multiple projects separately. So what he actually does is he gives each one of his projects, his own little notes and brain and memory so they don't just sort of overlap each other. We do a similar thing where we have multiple different projects, multiple different businesses running inside the memory galaxy. And so our agents, because it's so well organized and our agents actually update it automatically, they can sync between different projects. If we ask them about one thing and then another thing, they know how to tell the difference. And that's a great way to set up Obsidian. Now if you don't have Obsidian set up already, but you want a memory galaxy like this, basically you can see here, for example, we have a knowledge graph and everything is all like sort of segmented into different projects. So for example, we have people we work with. We have, for example, funnel strategy. Everything is a different markdown file, but it's all interlinked inside one knowledge graph. And I've never had to update it personally. My team does that for me. And so if you're wondering, okay, like how can you manage multiple different projects or different companies inside an operating system like this? Just build out Obsidian and make sure that you have everything separated but organized and tied together. So for example, one thing you'll see is that for each of our companies like, for example, the AI Profit Boardroom and Goldie Agency, number one, they have links between them. And number two, they're color coded, right? With bigger modules. So the good thing about that is, that is easy for our agents to visualize it and see the difference between what are our projects, who we're working with, etc.
For example, people are color coded differently to the way that we color code our projects. And this way, everything links together, but it's easy for your agents to organize between each other. Also, he's talked about multi-agent orchestration, which is something we've set up with Paperclip. So basically what you can have here is a team of agents working together. They can orchestrate tasks. We've got one CEO who basically in the org chart will command a team of agents who command their own sub teams of agents. And so the beautiful thing about that is it's easy for your agents. For example, if you got Claude or Hermes plugged into the system, it's easy for your agents to coordinate between each other and make sure that you can get the most out of every agent you have plugged in. We've got another question here from Daniel is asking about like animation, video agents and how do you get them working together? So one thing that you can do is you can use these open source projects. Like for example, we have open montage here. This can create kind of like films as you can see here, where you can basically in one single prompt generate a short feature film with AI and the quality of this stuff looks amazing.
So that's one way to do it. Another option is that you can use a video agent, but you tell it the specific style that you want. So if you want, for example, more like an animation style video or cartoon style video, then you would tell the video agent, okay, this is how I want it to look like.
Here are the characters I want to use, etc. And with the video agent, you could use that too. And also Sheena's given a really good tutorial here about how to use this as well and the models that you can use. So for example, you can use a combination of Kling, Seed Dance 12.7 to generate your agents together. And if you have a single mission control dashboard, with each tool having its own tab, like we have with the video and the open montage system here, it's easy to organize and coordinate them together. This is pretty wild. So Robbie shared agent operating system that he's built for lead generation. And this is basically using our agent operating system, but he's made it his own, right? So if you look at the system, basically what it can do is it can find potential prospects for contracting jobs and then reach out to those prospects once he's reviewed and approved it, right? So what it does is it finds the quick wins with the project details, then he approves the details, and then they're ready to send as an hourly system. And so he's basically got, he's called it the Deal Desk, which is amazing. It's basically a custom made system inside the Agent OS, where you can reach out and apply for jobs directly. So pretty insane. There was a really good share as well by Davide. So Davide was talking about Glm 5.2, which is pretty much up there when it comes to performing side by side versus Opus 4.8. But the difference is, of course, the number one is open source and number two, it is a lot cheaper on the coding plan. You get a lot more tokens. So if you've seen Glm 5.2, it's a really powerful system. We actually compared it side by side and outperforms Opus 4.8 on many of the tasks that we created. I mean, it can create some pretty insane stuff like check this out. This is a full open world game created by Glm 5.2. So it can create real stuff. It looks absolutely awesome.

6 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/1000774930769