Block open-sources Berd desktop application to unify enterprise AI agent workflows

Block has released Berd, an open-source desktop application designed to manage artificial intelligence agents across multiple models and tools. The software, available under an Apache 2.0 license for macOS, Windows, and Linux, reached its seventh public release on August 18.

The tool addresses a growing fragmentation problem in enterprise AI adoption. "We had capable agents through goose, Claude Code, and Codex, but working across them meant navigating different interfaces, configuration systems, and ways of managing context," said Lucinda Bell, Technology Communications at Block.

Berd provides a unified local environment where users can manage persistent projects, attach files, and configure various AI providers. "Berd is desktop-first because much of its value comes from working directly with projects, local files, tools, repositories, and agents running on or connected to the user’s computer," explained Brad Axen, Head of AI Capabilities at Block.

Rather than acting as a foundation model, the application functions as an orchestration layer. Constructed with Tauri 2 and React 19, the software communicates with Block’s open-source Goose agent framework via the Agent Client Protocol. Goose, which Block contributed to the Agentic AI Foundation in December 2025, runs as a bundled sidecar process.

For enterprise IT teams evaluating the software, Berd enforces strict version control. The build process pins a specific Goose backend version through a lockfile and verifies the cached binary, preventing the underlying agent runtime from silently drifting during updates.

For investors and corporate executives, the release highlights a strategic shift in the artificial intelligence market. The competition is moving from raw model capability to workflow orchestration, with Block attempting to establish open protocols as the default infrastructure for enterprise deployment.

This open approach positions Berd against proprietary alternatives like OpenAI’s Codex, Anthropic’s Claude Code, and Cursor. Unlike those tools, which are often tied to specific subscription models or aimed strictly at software developers, Berd is vendor-neutral and designed for broader business functions.

The application also introduces visual identities for different agents to help users distinguish between specialized tasks. Block designed these personas to make operational states visible, ensuring workers can track exactly which files, models, and instructions are actively shaping a conversation.

Portability and data control remain central to the architecture. Folders, repositories, and instructions are kept accessible outside the application, a local-first design that directly addresses enterprise concerns regarding data privacy and vendor lock-in in an increasingly autonomous AI landscape.