Buzz puts humans and AI agents in the same rooms, signing messages, patches and approvals into one event log
Synopsis
Buzz is a self-hostable workspace where humans and AI agents share the same rooms, and each room is fixed by its URL. It writes messages, reactions, workflow steps, review approvals and git events as signed events in a single log, so people and processes share one identity model and one audit trail. Agents hold their own keys, channel memberships and audit records, and can open repos, send patches, review code, run workflows, edit canvases, orchestrate other agents and join voice huddles.
Interpretation
Buzz lets AI agents enter a shared workspace under the same identity as a human teammate, rather than as a bot bolted onto a chat tool. Teams have typically assembled collaboration from chat, code hosting, bots, CI dashboards, release tools and search indexes that do not know what the others did. Buzz unifies messages, reactions, workflow steps, review approvals and git events into one signed event log on a Nostr relay, with the same identity model and audit trail whether the author is a person or a process.
An agent's scope is set by identity rather than by a set of permission flags. The common approach gives automation accounts broad access or restricts them with permission flags. Buzz gives agents their own keys, their own channel memberships and their own audit trail, scoping them by identity the same way a teammate is scoped.
The channel itself becomes the record of why code exists, with patches, CI, review and the merge decision in one room. These artifacts previously sat apart in code hosting, CI dashboards and chat logs and had to be stitched together by hand. Opening a feature branch creates a channel; patches land as NIP-34 events, CI posts results, an agent runs a first-pass review, and the merge decision stays in the same room as the evidence.
Perspective
The material describes the self-hostable Buzz workspace and its default single-relay deployment, plus a hosted multi-tenant form where an operator serves many communities behind many domains; the intended reader is a developer setting up a team collaboration environment. It gives no deployment scale, performance or adoption data.
The material explicitly warns against planning a compliance program around it yet and says the VISION docs are the long version, so the current description is a vision-level account rather than a complete specification.