Figma launches a built in AI agent for generating and editing full interface designs

Figma has launched a native AI agent directly inside Figma Design, integrating AI generated full screen designs into its collaborative interface design tool. The company had already added MCP integrations with tools like Claude Code and OpenAI's Codex, allowing users to turn prompted interface ideas into designs through external AI tools, but this new agent gives users a built in option inside Figma’s own workspace.

The AI agent works directly inside the multiplayer canvas, instead of relying on a separate floating prompt box or opening in another tab like Figma Make currently does, which is more focused on building functional prototypes and is not directly tied to design files from the main workspace. This new agent works on the selected design itself, so users can select an artboard and click the star icon in the top right corner to open a prompt box, where they can use natural language to generate new designs, update existing work, create interface elements, and automate repetitive design tasks. The generated results are also fully editable, since they behave like any other component.

Figma says multiple AI agents can run in parallel on the same canvas, with models fine tuned for design context and able to understand existing design elements. The agent is launching first in Figma Design, with plans to expand it to other Figma products like FigJam later.