Jack Dorsey, co-founder of Twitter and head of the fintech company Block, has introduced a new platform: Buzz. The free service aims to compete with Slack and GitHub by letting people work with AI agents on Buzz.
Jack Dorsey has unveiled a new, free collaboration platform called Buzz. The co-founder of Twitter and head of the fintech company Block announced the launch with a post on X.
There he describes the service as a “new group chat platform for teams and agents of all sizes, designed to reduce our reliance on Slack and GitHub.” The tool is model-independent, decentralized, self-determined and open source.
What is Buzz?
Buzz is a free, open-source workspace where humans and AI agents can communicate together in the same channels, share code repositories, and manage projects.
According to the official press release, the new service is based on the Nostr protocol. Similar to Slack, Buzz offers channels, threads, direct messages, voice features, media sharing, and automated workflows. The announcement states:
Over the last two years, we at Block have developed and open sourced a number of AI tools. We realized that the most productive work doesn’t happen when someone asks AI for help. Rather, it arises when humans and AI agents are in the same room, working on the same task and sharing the same context.
The crucial difference to a classic chatbot in a Slack channel: AI agents in Buzz not only respond to commands, but also have their own cryptographic identities, defined permissions and can post independently.
They can also review code, run shared automations, and participate in conversations. All processes take place in collaboration with human colleagues in the same work area.
“Every company will need a place where people and agents work together,” explains Bradley Axen, Head of AI Capabilities at Block. “The only question is whether this place is proprietary or open. We built Buzz because we believe it should be open.”
These models and tools can be integrated
Buzz is designed to be model and agent independent. Teams can use agents that run on any language model, such as Claude Code, Codex or goose. They can also bring their own agents or build completely new ones within the platform.
In addition, the new service can be connected to existing company tools, such as databases, CRM systems, code bases or file systems. Agents in the workspace can access it, depending on the permissions that each team configures.
Buzz is currently available as a desktop app for macOS, Windows and Linux. The tool is provided under an Apache 2.0 license and is available on GitHub.
Additionally, anyone can run their own instance on their own infrastructure or sign up through the website and create a community with hosting provided by Block.
Buzz has these advantages
The practical benefit of Buzz lies primarily in its bundling: Instead of maintaining chat, ticket system, code hosting and individual AI assistants as separate tools, these areas come together in a common workspace in the new tool.
Teams can switch directly from an ongoing conversation to planning, programming and project management without leaving the application.
The open source code also benefits development teams: Anyone who has their own requirements can adapt the platform to existing workflows or add missing functions themselves instead of waiting for an update from the provider.
This can be an argument, especially for younger companies and teams that are increasingly building their processes around AI agents.
Development status and similar projects
The platform is not quite finished yet: On its own website, Buzz expressly describes itself as being in the “early stage” and explicitly invites you to test it. The Git integration is also still in an early phase.
Dorsey is not alone with his business idea. Loud TechCrunch Paradigm partner and CTO Georgios Konstantopoulos recently introduced a similar open source product called Centaur, which he describes as a “virtual employee” and can be used either directly in Slack or via its own API.
Slack itself is also increasingly positioning itself as an agentic operating system through which AI agents can collaborate and coordinate.
In a discussion Hacker News A user who identified himself as a Slack employee expressed concerns about data protection: When multiple people and agents share the same channel, complex rules are needed to prevent agents from passing on information to people for whom it is not intended. This problem is much easier to solve with agents who only act on behalf of a single person.
Also interesting:
