Figma Unveils Design Agent

Figma has launched a Figma-native design agent that works directly on the canvas. Going beyond a simple generation tool, this agent deeply understands components, tokens, and standards within design systems to produce results optimized for actual editing work.

Key Features and Characteristics

• Canvas-embedded collaboration: You can start a prompt from anywhere on a layer without switching to a separate tool, and users and the agent can edit simultaneously.
• Parallel Prompting: Multiple design ideas can be generated and compared simultaneously for a single problem.
• Automating repetitive work: It quickly handles tedious busywork such as batch-renaming variables, swapping components, switching to dark mode, and filling in content.
• Exploring design directions: It supports 'Go wide,' which quickly generates a variety of styles, and 'Go deep,' which refines details while maintaining the design system.

Workflow and Integration

• Figma Make integration: After clarifying design intent with the agent, you can increase work efficiency by converting it into code via Figma Make and repeating the process in reverse.
• MCP server utilization: It works complementarily with the MCP server that handles workflows between code and canvas, maintaining synchronization between design and code.

Launch Information The beta will roll out gradually over the coming weeks to Full seat users on the Professional, Organization, and Enterprise plans. AI credits will not be consumed during the beta period.