Figma: FigJam is now your coding agent's whiteboard too

Figma published this announcement on April 28, 2026, alongside new MCP skills that extend what coding agents can do with FigJam boards.

The core addition is figma-use-figjam, a new MCP skill that gives agents direct read and write access to FigJam boards. Previously, AI coding agents could generate code and interact with design files in Figma, but FigJam — used for planning, diagramming, and team coordination — was outside that scope. Now agents can create and modify board content programmatically, not just generate static images of diagrams.

Two workflow skills launch alongside it. generate_diagram has been extended to handle complex architecture layouts and ERD (entity-relationship diagram) structures, which are typically the kind of content that developers produce before writing code. generate-project-plan converts specifications, codebases, or conversation threads into structured visual boards — project plans, milestone maps, or onboarding documentation — without requiring manual setup.

Figma describes the intended use case as closing the gap between code generation and team coordination. When an agent builds or refactors a system, it can now also document the architectural decisions visually in FigJam, making the reasoning available for engineering review before implementation proceeds. Teams can then reorganize, annotate, and build on that content collaboratively. The output from a FigJam session can also be exported back into the coding environment via the get_figjam tool.

The feature works with the Figma MCP server and is compatible with Augment, Claude Code, Codex CLI, Copilot CLI, Cursor, Factory, and Warp. It is available immediately in open beta at no cost, with Figma planning to move it to a paid API once they determine how agentic usage patterns map to seat-based pricing.

For design-adjacent teams — particularly product and engineering groups that use FigJam for planning — this means AI-assisted documentation of system architecture is now part of the same toolchain as code generation. Whether that tightens the collaboration loop or creates a new surface to maintain will depend on how teams integrate it into existing review practices.