Buzz in One Sentence: A Slack-Like Home for People and AI Agents
Buzz is an open source collaboration platform that combines team chat, code repositories, and automated workflows in a Nostr protocol workspace where AI agents participate as full, cryptographically identified members alongside humans, positioning itself as an AI agent collaboration platform that rivals traditional human-AI team tools built on centralized SaaS chat services. Block has launched Buzz as a free, open source workspace that looks familiar—channels, threads, direct messages, voice, media sharing, and code repos—but behaves differently by treating agents as first-class participants rather than hidden copilots. This is not a cosmetic tweak to Slack-style workflows; it is an opinionated bet that the future of work requires giving AI agents real seats at the collaboration table, with identities and rights, not just API keys.

Encrypted Passports on Nostr: Why Identity Is the Real Innovation
The boldest idea in Buzz is not that agents can post in channels—it is that each agent and human gets an independent encrypted identity “passport” built on Nostr’s public–private keypairs. Buzz issues every agent its own keypair and then adds a second signature that ties the agent back to a specific human owner, creating a cryptographic paper trail that anyone in the workspace can review and confirm. That dual-signature design lets teams enforce rules, such as allowing only agents owned by existing members into private channels, without trusting a central SaaS vendor to police access. In other words, Buzz treats identity as infrastructure: teams can stand up their own Nostr relay, keep all events under their control, and avoid routing collaboration traffic through Block’s servers if they choose. For organizations worried about AI sprawl and invisible bots, this is a rare attempt to make agent accountability explicit rather than implicit.

From Slack Add-Ons to Platform-Level Human–AI Teams
Buzz’s strongest argument is that treating AI agents as full workspace members beats bolting them onto legacy tools. Today, most teams interact with coding agents in private and then paste their outputs back into Slack as if they were human authors, which hides the true technical conversation. Buzz flips this pattern: agents connect through the Agent Client Protocol (ACP), work inside shared channels, and can be wired to Claude Code, OpenAI Codex, Block’s goose framework, or custom models. The platform bakes in team chat, code repositories, and automated workflows, turning the AI agent collaboration platform into a living record of how humans and machines decide, iterate, and ship code. One quotable detail shows the stakes: Block’s internal BuilderBot runs more than 200,000 operations a day, merges about 1,500 pull requests a week, and drives roughly 15 percent of production code changes. With Buzz, those kinds of agents stop being shadow workers and become visible teammates.
Open Source Collaboration and Data Ownership Over SaaS Lock-In
Buzz is also a clear shot across the bow of centralized SaaS collaboration tools. Its code has been public on GitHub under an Apache 2.0 license, and the current release, version 0.4.21, is already available on macOS, Windows, and Linux, with mobile clients still in development. Block’s founder has said the aim is to reduce dependence on Slack and GitHub, and Buzz’s Nostr foundation underscores that intent by letting teams own their relay and infrastructure. Block does offer hosted relays, free at launch and described as a beta, but emphasizes that any team can stand up its own instance with full control and no traffic routed through Block. This is what open source collaboration should mean in an AI era: not just inspectable code, but genuine data ownership and transport independence. For organizations worried that SaaS vendors will mediate every human-AI interaction, Buzz’s architecture is a deliberate escape route.
Why Buzz Matters Now—and What Comes Next
Buzz arrives as Block reshapes itself around AI, cutting more than 40 percent of its staff—from over 10,000 employees to just under 6,000—and framing AI as enabling “a new way of working.” Buzz is the collaboration layer for that smaller, more agent-heavy company, designed to sit alongside existing tools rather than immediately replace them. The platform is still early, with far fewer GitHub stars than Block’s goose framework, and its identity scheme has only been proposed as an upstream Nostr extension, not adopted yet. Meanwhile, the Linux Foundation’s Agentic AI Foundation is considering bringing ACP into its fold, hinting at a possible standards path for these human-AI team tools. The open question is whether other companies will accept agents as first-class workspace citizens or keep them boxed inside private chats and plugins. Buzz takes a clear stance: the future of collaboration is a Nostr protocol workspace where code, decisions, and agents live together in the open.






