Unlike traditional AI assistants that respond to individual commands, agents in Buzz possess their own cryptographic identities and permissions, allowing them to participate in workflows, review code, and contribute to conversations alongside human team members.
By utilizing the Nostr protocol, Buzz ensures that identities for both humans and agents are portable and verifiable, independent of any specific vendor or platform.
The system is model-agnostic, allowing teams to deploy agents powered by various large language models or frameworks such as Claude Code and Goose.
This infrastructure is intended to prevent vendor lock-in and provide organizations with full control over their data and AI integrations.
The platform is released under the Apache-2.0 license and is available for self-hosting or through a hosted version provided by Block.
Early features include integrations for databases, CRMs, and code repositories, with plans to expand Git capabilities to support open-source communities.
Block’s goal with the project is to establish an open standard for human-agent collaboration, ensuring that the workspaces where organizations build their AI infrastructure remain transparent and interoperable.