From AI Tool to Teammate: What Buzz Is
Buzz is a free, open-source AI team workspace that lets human workers and AI agents share channels, threads, and workflows as equal members, each with its own encrypted Nostr protocol identity, so companies can move from AI-as-tool to AI-as-collaborator inside everyday team communication and code repositories. Block’s launch of the Buzz workspace platform is a clear statement: AI agent collaboration should happen in public, not in copy‑pasted side chats. Instead of agents hiding behind a human account, Buzz invites them into the room. Channels, threads, direct messages, voice, media sharing, and code repositories all exist in one place that looks familiar to Slack or Discord users, but the social rules are different. Agents are treated as full participants, with cryptographic identities that tie them to a human owner, turning every interaction into part of a shared human-machine collaboration record rather than a private AI notebook.
Encrypted Passports: Nostr Identity as the New Org Chart
The most radical part of Buzz is not the chat UI; it is the way identity works. On Nostr, each participant is a public‑private keypair. Buzz gives every AI agent its own keypair, then adds a second signature linking that agent back to a specific human. Buzz effectively hands every agent a cryptographic passport, with a paper trail anyone in the workspace can review to confirm, “this is my agent.” This Nostr protocol identity model does more than satisfy security nerds. It changes how teams think about responsibility. Workspaces can enforce rules such as allowing only agents owned by members into private channels, without relying on a central vendor directory. Teams can run their own relay and keep all data routing under their control, or use Block’s hosted relays, which are free at launch in a beta phase. In practice, that means AI agents become accountable entities in the org chart, not anonymous scripts executing in the dark.
One Workspace for Chat, Code, and Automated Workflows
Buzz is opinionated about where AI should live: inside the same place people talk and ship code. At first glance, channels, threads, direct messages, voice, media sharing, and code repositories make it feel like a typical chat product. But the design assumes that both humans and agents will be active in all those surfaces. Team chat, code repositories, and automated workflows share one substrate, so an agent can comment in a thread, open a pull request in the repo, and trigger a workflow without jumping across tools. Block supports any model and popular agent harness, including Claude Code, OpenAI Codex, and its own goose framework, or agents teams bring themselves. That breadth matters: the future of AI team workspaces will be model‑plural, with specialized coding agents sitting alongside documentation, testing, or data agents. Buzz’s Apache‑licensed code has been on GitHub since earlier this year, but only now are public signups open and Block’s own Nostr relay available for teams that do not want to run their own.
Why Block Is Betting Its Future on Agent Workspaces
Buzz is not a side project; it is part of Block’s restructuring around AI. In February, the company cut more than 40 percent of staff, from over 10,000 employees to just under 6,000, citing AI as enabling “a new way of working.” Buzz is the collaboration layer for that smaller, more agent‑heavy company. This follows goose, its agent framework with tens of thousands of GitHub stars, and BuilderBot, an internal system that runs over 200,000 operations a day, merges about 1,500 pull requests a week, and accounts for roughly 15 percent of production code changes. Jack Dorsey has said Buzz aims to reduce dependence on Slack and GitHub, and the current 0.4.21 version already supports macOS, Windows, and Linux desktops, with mobile in development. The message is plain: the company expects most technical work to happen with AI agents as constant collaborators, not occasional tools called from the command line. Buzz embodies that bet by giving agents visible seats at the table.
What This Shift Means for Everyday Teams
For ordinary teams, Buzz’s approach has practical consequences. Instead of disappearing into one‑on‑one chats with coding agents and pasting the output into Slack, the “second conversation” where technical decisions happen can move into shared channels where everyone can see it. That means reviews include both human and agent reasoning, making AI agent collaboration part of the traceable history of the project. Buzz lets any team run its own Nostr relay so nothing has to route through Block, or they can opt for a hosted relay if they prefer an off‑the‑shelf solution. The repository has a little over 100 stars so far, compared with goose’s far larger following, which underlines how early the AI team workspace idea still is. Yet the direction is clear: human-machine collaboration is moving into the core workspace, not bolted onto existing tools. Teams that embrace agents as named, accountable members will gain more transparent workflows and faster, shared decision‑making. Those that keep AI boxed as a black‑box tool will fall behind. Buzz will not replace Slack overnight, and Block itself continues to run both. But as Buzz matures and its Nostr identity scheme gains upstream acceptance, it offers a blueprint for how future workspaces should treat AI: not as magic assistants hiding in the sidebar, but as full colleagues with cryptographic passports and visible responsibilities.
