different-ai/openwork

What it solves

OpenWork provides a centralized way to share and reuse AI workflows, skills, and Model Context Protocol (MCP) connections across different AI agents, teammates, and machines. It eliminates the need to configure the same tools and services (like Google Workspace or Microsoft 365) repeatedly for every individual AI tool you use.

How it works

OpenWork operates as a desktop application and a remote MCP server. By adding the OpenWork MCP to an agent (such as Claude Code, Cursor, or Codex), the agent gains access to a unified set of capabilities. It exposes two primary tools: search_capabilities to find available skills and execute_capability to run them. For organizations, the "OpenWork Den" serves as a control plane to manage access, provision inference, and publish skills to specific teams or users.

Who it’s for

• Individual users who use multiple AI agents and want a single source of truth for their AI skills and connected services.
• Organizations that upscaling AI adoption by managing shared capabilities and model provider access across a team.

Highlights

• Cross-agent compatibility: Works with any MCP-compatible client including Claude Code, Cursor, and Codex.
• Unified Workspace: Share skills, plugins, and MCP connections across different tools and machines.
• Enterprise Control: Admin interface for managing user access, desktop policies, and model provider provisioning.
• Plugin Import: Ability to import Anthropic-compatible plugins to make their skills available via the OpenWork MCP.