Before your team and its agents share one workspace.

The shared workspace is where a team's decisions get made and recorded: channels, messages, files, and increasingly the agents working alongside people. Buzz is one option in that category, and two properties separate it from the familiar tools. It is self-hostable, so the business can run it on infrastructure it controls rather than on a supplier's platform. And it is built on Nostr, an open protocol, rather than on one company's account system.

Neither property is an advantage by itself. Both move responsibility from a supplier to the business, and where responsibility lands is the entire question worth asking.

• Identity is a cryptographic key pair rather than an account on a supplier's server. No administrator can reset it for you, and whoever holds the private key holds that identity until the key is replaced
• Messages are signed events published to relays: independent servers that store and serve them. Which relays a workspace uses decides where its record actually lives
• Deletion is a request to relays rather than a guarantee. A relay may honour it, and a relay that already served an event to someone cannot unsend it
• Self-hosting moves backup, upgrade, incident and recovery duties to whoever runs the server. That is real control, and it is also work that a named person has to own

• The workflow: one real process, what it costs today in time and errors, and the outcome that would count as an improvement
• Channel and membership design: who can see what, and what changes when somebody joins or leaves
• Agent identity and permissions: what each agent may read, may recommend, and may do without a person approving it first
• Relay and hosting responsibilities: which relays hold the record, who operates them, and who is called when one stops responding
• Data sensitivity and retention: what may enter the workspace, what must never, and how long each kind of thing stays
• Integrations and connector gaps: what genuinely connects to the systems already in use, and what would have to be built
• Audit, backup and recovery: export the record, restore it somewhere else, and confirm what came back is complete
• Migration from email, WhatsApp, Slack or Teams: what moves, what is left behind, and who tells customers where to reach you
• Support, upgrade and incident ownership: named people, agreed response expectations, and a path that works when something breaks
• Cost and success measures, written down before a trial starts rather than argued about afterwards

Buzz should be treated as an evolving platform. The deployed version, approval enforcement, export, retention, security and operational behaviour must be tested before anyone relies on them contractually.

It is not an unconditional replacement for every Slack or Teams workflow, and no workspace choice creates compliance, security or data-location guarantees on its own. Not every business needs a different workspace: the honest starting question is whether the current one is actually causing the problem.