Jack Dorsey’s Block open-sources Berd — a desktop workspace for AI agents that keeps conversations local

Block, the technology company founded by former Twitter CEO Jack Dorsey and owner of Square, Cash App and Tidal, has open-sourced Berd — a desktop application originally built for its own employees to work with AI agents across different models, tools and projects in a single environment.

Berd is a locally installed graphical desktop application, not a browser-based service. It is released under the permissive Apache 2.0 license, meaning anyone can use, modify and redistribute it, including for commercial purposes. Free downloadable builds are available for macOS, Windows and Linux. The GitHub repository has reached version 0.6.2 as of August 18, its seventh public release, and lists 91 contributors.

Block positions Berd as a "daily AI work surface" — a single place where users can start chats, attach files or folders, choose agents and models, work inside persistent projects, configure AI providers, manage skills and extensions, review session history and build automations. The application stores all conversation history locally on the user's machine and does not send it to Block's servers.

Berd supports multiple AI providers: OpenAI, Anthropic, Google, Groq, Ollama and OpenRouter, along with any models compatible with them. Users can switch between different models within a single conversation. The application includes "skills" — extensions that allow agents to perform actions such as working with the file system, executing Python code, reading PDFs, searching the web, and interacting with Git. Block has also released an SDK for building custom skills.

Brad Axen, Head of AI Capabilities at Block, told VentureBeat that Berd was designed as a desktop-first application because much of its value comes from working directly with local projects, files, tools, repositories, and agents running on or connected to the user's computer. The company believes that keeping data local is a key advantage for privacy and control, especially for developers and power users who need to work with sensitive or proprietary code.