Buzz from Block: Open-Source Slack for Humans and Agents
Your team's work is spread across chat, the code forge, CI dashboards, release tools, and a pile of bots glued together. Your agents only ever see a slice of it, and tracing back what one of them did means digging through every one of those tools. Buzz from Block puts all of it into a single event log you host yourself.
It is fully open source, and it treats agents as real teammates: their own identity, their own channel access, and their own audit trail.

What Happened
Buzz is an open-source workspace where humans and agents build together. Everything that happens on a project, from every message and reaction to every patch, review approval, and workflow step, is written to one event log as a signed event. People and agents write to that same log the same way, each with their own keypair.
Once an agent is inside, it can do what a teammate does. It can open repos, send patches, and review code. It can run workflows triggered by messages, reactions, schedules, or webhooks. It can create channels, edit canvases, and pull in the right people. And it can search six months of project history and post the matching threads as evidence.
A feature branch becomes a channel, so the patches, the CI results, the agent's first-pass review, and the merge decision all land in that one room instead of scattering across five tabs. Buzz works with Claude Code, Codex, and Goose. You can run it locally, on your own server, or deploy a relay to Railway in one click. Adding an agent to a project works the same way as adding a person to a channel.
Why It Matters
If you already run coding agents against real repositories, the signed event log is the part worth your attention. It gives you one place to see what an agent touched, what it changed, and who approved it, without stitching together five dashboards after the fact. Because Buzz is open source and self-hosted, that record stays on hardware you control.
More shares you might like
ToolsHow to Make Motion Graphics Videos With ClaudeAlex Ruiez · 1d · 33 views
NewsNVIDIA Open-Sources Auto Ontology, a Semantic Layer for AI AgentsAlex Ruiez · 2d · 46 views
ToolsHow to Write Specs Coding Agents Can Build Without GuessingAlex Ruiez · 3d · 50 views
AIParallel's Decisions API Grounds Decision Models With Live Web SearchAlex Ruiez · 4d · 55 views
How to Claim 1B Muse AI TokensAlex Ruiez · 4d · 103 views
ToolsHow to Use Rich Clicker Macro's Smart Auto ClickerAlex Ruiez · 5d · 59 views
Comments
Join the conversation — sign in to comment.
No comments yet — start the conversation!