figma guide
Figma May 2026 Updates: AI Agents, Custom Skills, Mobile Make & More
A practical breakdown of Figma’s May 2026 updates: MCP AI agents, Custom Skills in Make, mobile Make previewing, ChatGPT Images 2.0, Draw improvements, and performance gains.
- Published
- Updated
- May 13, 2026
- Read time
- 7 min
Figma’s latest round of updates for May 2026 is one of the most significant releases in recent memory. From AI-powered agentic workflows to a massive performance overhaul and new Draw capabilities, there’s something valuable for every designer. Let’s break down what’s new and how these changes affect your workflow.
MCP Server: AI Agents Can Now Work Inside Figma
One of the biggest announcements is Figma’s official Model Context Protocol (MCP) server, which lets AI agents retrieve context from your Figma files and take actions directly on the canvas. This opens the door to truly automated design workflows.
Three pre-installed MCP skills ship with the update:
/prototype-to-figma— Takes a running local prototype and converts each unique screen into a Figma frame automatically./figma-generate-designand/figma-generate-library— Capture design tokens and variables from your codebase and sync them into Figma./figma-use— Allows an agent to create and modify a real Figma design file using your existing components, with full write access to the canvas.
The MCP server supports both remote and desktop deployment, making it flexible for solo designers and enterprise teams alike. Combined with Code Connect, agents can now keep design systems synchronized between your Figma files and the live codebase.
Figma Make Gets Smarter and More Controllable
Figma Make — the AI-powered prototyping and build environment — received a cluster of meaningful improvements on April 30:
Voice-to-Text Input
You can now dictate prompts directly into the Make chat interface. Your speech is transcribed and cleaned up before you review it, making it faster to describe complex changes without typing.
Question Cards
Instead of open-ended prompts, Make now surfaces structured question cards at decision points, giving you guided options rather than a blank slate. This is especially useful for designers new to AI-assisted workflows.
Version History
Make now tracks changes as you iterate, letting you revert to any prior state without losing work. This was a heavily requested feature and makes experimental workflows much safer.
Zapier Connector
A new Zapier integration brings real-world data directly into Make sessions — pulling content from Google Drive, Microsoft Office, Zoom, and 9,000+ other connected apps. Populate your designs with actual content instead of placeholder text.
Clear Context Button
A one-click reset clears the current session context, letting you start a fresh AI conversation without losing your canvas work.
Figma Draw: More Powerful Vector Editing
The Draw mode received a major quality-of-life update on April 29:
Auto Layout in Draw
You can now apply auto layout directly in Draw mode without switching to Design mode. This removes a frustrating context switch that slowed down vector-heavy workflows.
Inline Layer Type Labels
The layers panel now shows component, instance, and text layer types labeled inline, so you can tell at a glance what kind of layer you’re looking at without expanding or hovering.
Text on a Path
A new dedicated tool lets you place text along any existing path or generate text on a circle. This brings a feature that designers have long worked around with workarounds into a first-class, purpose-built tool.
Enhanced Brush Controls
Brushes now support gradient and blend mode adjustments, enabling richer texture work and more precise, expressive vector illustration directly within Figma.
FigJam: Better Collaboration Tools
FigJam’s April 28 update focused on real collaboration improvements:
- Cell merging and text color in tables, bringing FigJam tables closer to what teams expect from a whiteboard tool.
- Improved arrow rendering for cleaner diagrams.
- Shape-flipping via drag handles — no more right-click menus for simple transforms.
- Template publishing for Professional plan users, allowing teams to share reusable board templates across the organization.
FigJam + AI Coding Agents
Also on April 28, FigJam gained the ability to generate architecture diagrams and entity-relationship diagrams directly from coding agents via updated MCP tools. This means your AI coding assistant can now push visual diagrams into FigJam as part of a planning or documentation workflow.
Performance: The Numbers Are Striking
Figma’s April 24 performance update deserves its own section because the improvements are dramatic:
| Area | Improvement |
|---|---|
| Vector editing speed | Up to 10x faster |
| Make prototype frame rate | 4x smoother |
| Memory warnings | Reduced by 92% |
| File load times | Significantly improved |
For designers working with large, complex files — dense component libraries, multi-page design systems, or heavy illustration work — these gains will be immediately noticeable. The reduction in memory warnings alone addresses one of the most common complaints from professional Figma users.
Desktop App Improvements (May 1)
The Figma desktop app on macOS also received practical quality-of-life upgrades:
- Tab search — Search across open files directly from the tab bar.
- Direct link opening — External links now open in your default browser instead of inside Figma.
- Background file preloading — Files start loading before you switch to them, making navigation between projects noticeably smoother.
AI Image Generation Improvements (April 29)
Adding image references in AI generation workflows got more flexible. You can now provide reference images through a dedicated button, copy-paste, or drag-and-drop into the prompt interface — and this works consistently across all Figma products, not just in specific contexts.
Custom Skills in Figma Make (May 11)
One of the most practical additions in mid-May is Custom Skills — markdown files that capture the conventions and workflows you repeat most often. Instead of re-explaining your design system rules or naming conventions every session, you define them once as a skill, import it into Make, and call it with a slash command from any prompt.
You can create skills from scratch inside Make or import them from an existing file. Figma has also signaled that team and organization-level skill sharing is coming soon, which would let design leads distribute standardized workflows across the whole team.
This is a significant step toward reproducible AI-assisted design — less prompt engineering per session, more consistent output across projects.
Figma Make on Mobile (May 11)
Make is now available on the Figma iOS and Android apps. You can preview a Make prototype directly on your phone, interact with it using real gestures and device constraints, and share it for review without leaving your mobile device.
For teams that design primarily for mobile, this closes a real gap. Testing on an actual device catches layout and interaction issues that desktop preview reliably misses. The ability to share for review from the same session shortens the feedback loop with stakeholders who prefer reviewing on their phones.
ChatGPT Images 2.0 Now Available in Figma (May 11)
OpenAI’s ChatGPT Images 2.0 model is now integrated into Figma’s image generation features. It is available in Make Image and Edit Image across Figma Design, Draw, Slides, Buzz, and FigJam — as well as in Figma Weave.
This is a direct upgrade to the AI image generation layer. The new model produces sharper, more prompt-accurate results than the previous generation, which matters most for mockups, placeholder images, and concept explorations embedded directly in live design files.
What This Means for Your Workflow
These updates represent a clear direction: Figma is becoming the hub where AI agents, design systems, and development workflows converge. The MCP server is the most significant infrastructure addition — it turns Figma from a passive design file into an active participant in automated pipelines.
The mid-May additions extend that direction. Custom Skills make AI-assisted workflows repeatable at scale. Mobile Make previewing brings device-accurate testing into the same tool. And ChatGPT Images 2.0 raises the quality ceiling for AI-generated visual content inside your design files.
The performance improvements, meanwhile, address longstanding pain points for power users. Faster vector editing and smoother prototypes mean less time waiting and more time creating.
For most designers, the practical impact will be felt in four areas:
- Faster iteration — Performance gains and Make improvements reduce friction throughout the design process.
- Better collaboration — FigJam updates, template publishing, and mobile Make previewing improve team workflows.
- AI integration — Whether you adopt agentic workflows immediately or not, Figma is laying the groundwork for a fundamentally different relationship between design tools and AI.
- Repeatable workflows — Custom Skills mean your AI sessions can carry your team’s conventions automatically, without prompt repetition.
Stay updated with Figma’s official release notes as the platform continues to evolve through 2026.
Related JustFigma guides
Keep exploring this topic with these related JustFigma resources:
§ Keep reading