Jack Dorsey's Block Launches Buzz, an Open-Source Workspace for Humans and AI Agents
Built on the Nostr protocol, Buzz puts team chat, AI agents, and Git code review in one Apache-licensed app aimed at Slack and GitHub.
Block's free, open-source app puts Claude Code, Codex, and human teammates in one chat window.
Block has launched Buzz, a free and open-source workspace that puts team chat, AI agents, and Git code review in a single interface, positioned as a direct alternative to Slack and GitHub.
What Buzz actually is
Buzz launched July 21, 2026 out of Block, the company Jack Dorsey runs alongside Square, Cash App, Afterpay, and Tidal. The app combines channels, threads, direct messages, shared canvases, media, search, an audit log, and YAML-based workflows, with a packaged desktop client for macOS, Windows, and Linux. Git hosting and code review sit inside the same interface as the chat, rather than living in a separate tool a team has to switch to.
The distinguishing feature is that AI agents are first-class participants rather than bots bolted onto a channel. Any agent that speaks the Agent Client Protocol, including Claude Code, OpenAI's Codex, and Block's own agent called goose, can operate inside a Buzz channel alongside human teammates, reading and writing to the same conversation and the same code review thread.
How it's built
Buzz is a Nostr relay written in Rust, released under the Apache 2.0 license. Every message, reaction, workflow step, code review comment, and Git event is a cryptographically signed event written to a single log. That's what lets the decentralized Nostr protocol carry chat, code, and agent activity through one channel instead of stitching together separate systems, and it means no single company's server has to be the source of truth for the conversation.
Why now
Dorsey's stated reason for building Buzz: startups increasingly depend on AI agents to get work done, but teams end up coordinating that work across different platforms with no shared thread connecting a chat message to the code change it produced. He's described the platform as model-agnostic, decentralized, self-sovereign, and open source, language aimed squarely at teams wary of being locked into one AI vendor's chat product or one company's hosted infrastructure.
Why a build studio cares
This is one of the first serious attempts to formalize how AI agents sit inside a team's actual working tools rather than as a separate chat window off to the side, and the Agent Client Protocol compatibility is the part worth watching. In principle, an agent a client already runs through Claude Code could show up as a teammate inside Buzz without custom integration work. For any client whose engineering team already runs more than one coding agent day to day, that's a real workflow question, not a speculative one.
Next step: read the launch coverage at TechCrunch or the technical rundown at Runtime Wire. If you want help wiring agent-native chat into your own team's stack, write to hello@gattyworks.com.