AI Pulse · Under the Radar

Jack Dorsey’s open-source Slack alternative, Buzz, quietly became something else this month. Users figured out they can drop Hermes Agent instances into Buzz channels and treat them like free AI employees. No new subscriptions, no extra API costs. If you already pay for Claude or OpenAI’s code models, you can spin up agent teams in about five minutes.

Two creators published walkthroughs within a week of each other, both landing on the same workflow. Riley Brown posted a 35-minute two-part tutorial showing how to stack agents inside Buzz channels, each handling a different task. Julian Goldie SEO followed with a nearly identical setup. Neither coordinated. They just stumbled onto the same pattern because the tooling made it obvious once you opened it.

What this actually means

You get a team chat tool that doubles as an agent orchestration layer. One agent drafts reports, another pulls data, a third reviews output. They hand off work in threads, same as human teammates. The cost sits with your existing API plan, not a new SaaS tier. Buzz itself is open-source, so you can host it or fork it if you want control.

This matters because most agent platforms sell you the orchestration, the hosting, and the models as a bundle. Buzz splits that apart. You bring your own models, run your own infrastructure if you choose, and the coordination layer costs nothing. It is not polished. It is not marketed. But it works, and people are building real workflows without waiting for a vendor to package it.

We track patterns like this inside the Omni Command Centre because the gap between “someone figured it out” and “your competitors are using it” collapses faster than most owners expect. If your team already uses Slack and already pays for Claude, this setup costs you nothing but setup time. That window does not stay open long.