Ruflo Agent Swarms: Automate ANYTHING! 🤯

So today we are going to be looking at Rufflow which is a free extension for Claude that allows you to deploy multi-agent swarms so you can coordinate autonomous workflows and build more powerful autonomous systems with multi-agent AI orchestration using Claude Code. What that essentially means is that you can just build a huge team of agents using Rufflow. Now if you want to set this up with Claude Code it's pretty simple you can just use the CLI instructions right here right so you copy these commands and then you would go into terminal and set them up right. Once you've done that then you can open up Claude like so open up Claude here and once we set up Claude then we can say okay are you using Rufflow and we'll just double check this is working and you can see here it says Rufflow is configured in our environment right.

So step number one you install it via CLI using these instructions from the github. Step number two make sure it's actually working inside Claude Code as you can see. Now you might be saying okay why would I use agent swarms or how can I use them? So the best thing that you can do is actually ask Claude okay based on what you know about me what are the best use cases for Rufflow agent swarms.

So let's try that out this is actually based on a question from Norman Vale inside the air profit boardroom who's saying like is anyone using Rufflow swarms for certain things at the minute I'm trying to set up an automation for this I'm also using it for researching stuff struggling with the most efficient ways to get that set up right and so whatever you're trying to do inside Rufflow you can just speak inside Claude Code and ask it for the best tips right. So you can see for example here here's some ideas and I'm actually going to connect this to my obsidian vault because it needs more context. For some reason Claude says I don't have any saved memory about you yet which is crazy because I use Claude every day. All right so now it's looked at my obsidian vault which is where I store my memories this is why it's really useful to have an obsidian or second brain.

I've got loads of training on this inside the airport for boarding but basically you want to really understand how to store your memories in the context of you just in case stuff like that happens and then you can see here it's given me five different workflows that we could use here. So for example it could be like SEO keyword and article briefing it could be like AIPB workflow updates it could be article generation right and then it could also be like content production pipelines right. So I'm going to go with method number four I'm going to say okay go with method four using Rufflow agent swarms and then you go from there. If you got any questions as well feel free to ask and I'll answer them at the end.

So you can see now it's pulling in some examples is what keywords you want so I'm going to say just use the AI agents automation niche and then where should the article briefs be saved so we can actually save them to obsidian vault which is what I'm going to do. Submit the answers here and now it's going to start creating that agent swarm. Now you might say why should you care about Rufflow like you've already got Claude set up. Here's what here's why Rufflow helps you right number one it automates tasks that you probably have a team of people doing so it can do a lot of human tasks directly for you.

Number two it learns from every task so it gets smarter over time and it just self-manages. You can see it's building out the agent team here with the swarm which is pretty cool so it's creating like for example a researcher for Claude, researcher free agents, researcher on automation, comparisons, ARCO etc. So it's building out a full team of AI agents as you can see here and then you can have multiple agents working in parallel not one by one so it makes it much more efficient as a process. It can remember things across sessions and it connects to tools you already use right.

So you see here it says six agents are now running in the background and it's created a list of all the agents that it's got running right here and now it's saying for example I'll notify you when the architect finishes the files and the files land in your ball right. So these agents are just working in the background and the agent SEO architect has consolidated the research into Obsidian now so you can actually write those files into your memory too. So if we go into Obsidian as you can see it's now organizing that information into a nice little test which is pretty cool and then you can see the progress here so we've got like the main these are the agents working on so far so we've got three local agents now working and it gives you the progress so far so one out of five tasks are complete. Now you might want to know okay how does Ruffalo actually work so you want to think of it like a restaurant kitchen there's a head chef which is the router and then there are specialist cooks the agents one for salads one for grills one for desserts and so you basically got a recipe book here right and so you can have multiple agents working on multiple tasks it's really good for research content creation and you basically have a swarm of agents that just go off and do their own thing here which is pretty powerful and so that's how to use it how to set up some of the most efficient ways to automate and you can just leave it running in the background like this.

The other thing that I would say here is when you're using this when you're using Ruffalo swarms with Claude you want to make sure these specifically tell it in the prompt to use Ruffalo otherwise what will happen is it will just use Claude directly for agent swarms which is not what you want and then you can see here the architect is running with all the five research sets completing it's written the briefs and save them directly to this folder locally and you can see the briefs being written by Ruffalo here right so it's actually creating the markdown files updating it in real time and it's writing the briefs for all the keywords that we're targeting which is pretty cool so this is Claude with Ruffalo just automated in the background you can see we've got three different briefs already we will need five but that's basically how you can automate anything with this which is pretty powerful so if you want the full guide from today I actually answered Norman's question on a video and any sort of questions like this I answer them with videos just to help the community as much as we can inside the airport for boarding so if you want the full guide it's inside the new tutorial section as you can see right here so we've actually got a full guide on Ruffalo how to use it a step-by-step operating procedure 30-day roadmap and just to recap Ruffalo is an AI agent orchestration platform that runs on topical code gives you 100 specialized AI agents each one built for a specific type of work agents work in swarms which are self-organizing teams it has persistent memory it can connect to all sorts of different other options and it basically gives you a full AI team that works together right and basically this helps you like create but you could have a hundred AI agents working together if you wanted to so we've put it inside the new daily update section here if you want to get your questions answered next time just post inside the community section and I create new videos for people daily inside the calendar you can jump a weekly coaching course you can get all my best AI training here including how to go from beginner to expert with AI automation inside the classroom and then you can also connect with me personally too but let's see what questions we got here oh we got a lot step fun is my favorite says intimacy yeah it's really good for that it seems pretty good and fast for using with Hermes and it's pretty cool that it's free as well what are some absolute must-do's for configuring Hermes agent I honestly I just go with the defaults right but if you wanted to play it super safe then you would host it elsewhere but I just don't like to I think the main thing is the absolute do and don't is don't give Hermes access to anything that you're not comfortable with right as long as you don't do that then you're okay anything that you shouldn't give access to don't do it and you could set it up on a VPS or that sort of thing but you have to be careful that you could silo it away on docker that's a good option but bear in mind like I've never heard of any serious issues with Hermes so far and it seems pretty good so far Damavicious says the most down-to-earth person thank you very much sir Jovi says I have to tell my Claude code CLI to log memory or I won't do it yeah that was weird I'm surprised Claude didn't have any memory with me I suppose I've built in persistent memory but I'm really glad that I had obsidian set up that helps a lot isn't this kind of irrelevant though because of Claude's new agent view that lets you run multiple parallel agents so this option right here Claude already had a setup previously where you could run agent teams in parallel but with Ruffalo you can create like a hundred agents in parallel so it's quite different talking about Ruffalo yeah how many skills you have I just use the built-in skills and then I created a few for workflows that are custom but yeah inside Hermes agent I just mostly default skills and then I write a few of my own not that many does setting up multiple agents with a tool like Ruffalo yes is it will use more tokens right you have to be careful with that so yeah be aware of that if you're running 100 agents are you going to use more tokens absolutely how can I do something like you do I have Hermes so check out all the tutorials and Hermes I have inside the AI Profitable Boarding everything that I personally use with Hermes is inside the AI Profitable Boarding here so we've got loads of tutorials video guides step-by-step breakdowns etc all my best trainings are inside the AI Profitable Boarding on that I don't know what to start with so if you're not sure where to start go to the classroom and then go to the beginner to expert section here you can go from there someone said I work at Anthropic we saw this was a big issue over the past couple of months but a few days ago we made massive improvements nice man that's awesome I got a Corey says I've got a couple of open core agents and a few things set up but this has to be decent too many are talking about it and they're all saying different use cases do you have Obsidian set up yeah as I showed earlier so we have Obsidian set up locally and then that links to any agent that I directed to Memplace versus Obsidian I found Obsidian is just like the simplest easiest way to give access to my agents context so I found it's a lot easier it's a lot more chill all right thanks so much for watching