Figma Made Its Design Tools More Accessible To AI Agents

Figma made its design tools more accessible to AI agents with major updates to its Model Context Protocol (MCP) server and Figma Make AI app builder. The move is set to transform how developers, designers, and AI coding agents collaborate by bridging prototypes with real, editable code.

Figma’s MCP Server Opens Doors To AI Integration

At the heart of this update is Figma’s MCP server, a powerful bridge that allows AI models to directly access the code behind Figma designs. Previously, AI tools could only interpret rendered prototypes or images. Now, with MCP expansion, they can view the underlying code of a Figma Make file, eliminating guesswork and enabling more accurate builds.

Figma’s Chief Technology Officer, Kris Rasmussen, explained that the MCP server indexes code in Make files, so AI platforms can request exactly what’s needed. This creates a smoother workflow for AI-powered coding tools and agents.

Why This Matters For Developers And Designers

By making its design tools more accessible to AI agents, Figma is creating new opportunities for both developers and designers. Instead of relying on static mockups, AI can now read, recreate, and extend designs with far greater accuracy.

The MCP server now works with products from Anthropic, Cursor, Windsurf, and VS Code. Remote access also means developers no longer need to be tied to Figma’s desktop app, opening the door for more flexible, cloud-based collaboration.

A Step Toward AI-Native Design Workflows

This update signals Figma’s vision of AI-native design workflows. With AI agents able to directly interact with code, designers can focus more on creativity, while AI assists with execution. It’s a move that aligns with the growing trend of AI-assisted development across industries.

Figma also hinted at upcoming features like Design Snaps, suggesting that more AI-enhanced tools are on the horizon. By empowering both coders and creators, Figma is positioning itself as a central hub where design meets AI automation.