Hermes Agent Kanban Swarms are INSANE!
Hermes Agent released a major update enabling concurrent multi-agent Kanban operations without freezing. The update allows teams of agents to work in parallel on tasks across multiple boards, exemplified through a content factory that auto-deploys SEO articles to websites.
Summary
The Hermes Agent platform introduced a significant update that fixes the fundamental bottleneck of running multiple agents simultaneously on Kanban boards. Previously, the system would freeze when multiple agents tried to save work to a shared database because only one agent could write at a time, forcing others to wait. The update implements a "back office helper" system for slow database saves, allowing agents to continue work without blocking each other.
The speaker uses an analogy of 10 AI workers sharing one assistant's notebook—the old system forced all workers to wait when one worker's entry was being saved. The new system delegates slow saves to a background handler, enabling true parallelization. This eliminates the previous limitation where running more than one or two agents would break the entire system.
The practical demonstration shows a content factory implementation where specialized agent profiles (SEO writer, SEO backlink builder, SEO keyword generator, SEO outline writer) work collaboratively across multiple Kanban boards. A single task is automatically triaged by a head agent into different workflow stages (triage, ready, done). The system successfully created and auto-deployed five SEO articles to a live website without manual intervention.
The speaker outlines an "unfreezable loop" methodology with five principles: identify what's freezing agents, move slow jobs out of the way, keep teams moving, prevent duplicate work, and lock the system to prevent breakage. The update enables horizontal scaling across multiple boards with different projects and agent teams, while token usage can be optimized through free local models or alternative APIs. The broader Agent OS system integrates this Kanban functionality with additional features like the Hermes Oracle for news, voice-activated agents, media generation, and lead generation workflows.
Key Insights
- The previous Hermes system would freeze entirely when running more than one or two agents simultaneously because database saves were sequential, forcing all agents to wait when one agent's save operation was slow.
- The new update implements asynchronous database saves using a background helper system, allowing multiple agents to write to the database without blocking each other, eliminating the primary bottleneck for multi-agent parallelization.
- A single Kanban board can now orchestrate multiple specialized agent teams working on different tasks simultaneously—the speaker demonstrated five SEO articles being created and auto-deployed to a website in parallel rather than sequentially.
- The system architecture uses a head agent for automatic task triage and routing to specialized sub-agents across multiple boards, similar to how Claude Ultra uses an orchestrator agent delegating to sub-agents.
- Token usage concerns can be mitigated by using free local models like GLM 5.2 or free APIs instead of paid models, making large-scale multi-agent operations economically feasible.
Topics
Transcript
[0:00] Today, we have a brand new update from Hermes Agent on Kanban boards. So, this is a brand new update that basically means that you can build more and better multi-agent Kanban teams. And I'll show you some examples of what we've done with this in a second. And you can have larger concurrent sessions and more with this new update. Now, you can see the tweet right here. I'm going to break down exactly what this means step by step so that if you're not sure how to use Kanban boards or if you're [0:30] wondering, "Okay, what are the actual use cases that are in teams of agents orchestrated with a Kanban board?" Let me show you…
Full transcript available for MurmurCast members
Sign Up to AccessMore from Julian Goldie SEO
How to Run DeepSeek V4 Flash for FREE!
A tutorial demonstrating how to use DeepSeek V4 Flash for free through Open Code and integration with agent operating systems like Hermes Agent. The speaker showcases building websites and apps using this free AI model and explains how it compares favorably to larger models despite being smaller.
Microsoft Fara1.5 27B NEW Browser Automation Model is WILD!
Microsoft released Phi-3.5, a family of three computer use models (4B, 9B, 27B) that automate browser tasks through vision-based clicking rather than HTML parsing. These open-weight models significantly outperform larger closed-source alternatives like OpenAI's Operator and Google's Gemini 2.0 on web automation benchmarks.
Claude Obsidian 2.0 is INSANE (FREE!)
Claude Obsidian 2.0 is presented as a free AI memory upgrade that allows users to upload files into a folder for permanent retention and linking. The system creates a knowledge graph that learns from business documents, provides sourced answers, and can be shared across teams.
Claude Agent OS is INSANE! 🤯
Julian presents a comprehensive Claude-based agent operating system that integrates multiple AI models, automated workflows, and a persistent memory system to automate daily tasks. The system runs 24/7 and uses free or existing subscriptions, combining tools like voice agents, content creation, competitor monitoring, and real-time news analysis into a single unified dashboard.
NEW ChatGPT Update is INSANE!
OpenAI released a major ChatGPT update featuring a Chrome extension called Side Chat and an improved desktop app that work together to streamline SEO research and content creation. The update allows users to analyze multiple browser tabs simultaneously, highlight text for quick answers, and convert research into finished work without constant tab switching.