Figma launches a native AI agent on the canvas
The agent rolls out gradually with a waitlist open now, sitting alongside (not replacing) Figma's MCP server.
Figma has launched a native AI agent that lives directly on the canvas and in the left rail, rather than as a separate tool. Figma says it's fluent in the app itself, with context on a team's components, tokens, standards, and best practices that isn't available to third-party tools. It's rolling out gradually over the coming weeks, with a waitlist open now.
Because the agent sits on the canvas, there's no separate setup or context switching: prompts can start from any layer, several ideas can be explored in parallel, and edits can be made by hand while the agent works alongside. It's positioned as separate from Figma's MCP server, which is meant for moving work between code and canvas, while the agent is for working directly in a file.
- Design exploration: generate multiple stylistic directions or layout options from a single prompt, using a team's most frequently or recently used components as a default, or steering it by picking a library or @ mentioning specific tokens, variables, and components.
- Bulk edits: update typography across a file, replace placeholder copy and imagery across a grid, set components to a specific state, or convert screens to dark mode without manual fill and contrast changes.
- Design system upkeep: bulk-update descriptions, tags, and use cases across libraries, standardise naming conventions, and document components with their states and variants.
