Penpot is pushing the boundaries of design workflows by experimenting with MCP (Model-Driven Communication Protocol) servers to enhance AI-powered design processes. Unlike traditional AI models that can't interact with design files, Penpot's MCP servers can understand user intent and translate that into structured API requests. This capability allows for tasks like exporting only the icons used in a design. The servers act as a secure bridge, ensuring that user data remains private while still enabling advanced interactions with AI. The potential applications of MCP servers are vast. Penpot’s approach allows designs to be expressed as code, facilitating seamless transitions between design and code. This means users can convert design elements into code and vice versa, generate documentation, and even create new components based on existing design systems. Examples from demos show how these servers can automate complex tasks, such as translating a Penpot board into production-ready HTML and CSS or conducting visual regression testing. Penpot is still in the experimental phase with these servers and is actively seeking beta testers to refine their functionality. Users have already started exploring various use cases, demonstrating the versatility of MCP servers in streamlining design workflows. The initiative aims to foster collaborative development of tools that meet the needs of designers and developers, providing a platform for innovative ideas and improvements.