Starting today, work with an agent that is built for Figmaβdirectly on the canvas.
Share The Figma design agent is here
Designers need purpose-built tools that serve the essentials: exploration, experimentation, collaboration, and precision. Figma was built as a multiplayer canvas to make all of that possible. As teams adopt agentic tools to build products more quickly, false choices are emerging: Speed or precision? AI generation or direct manipulation? You shouldn't have to choose.
Earlier this year, we opened the Figma canvas to third-party agents so you could bring design system knowledge into your agentic workflows. Now we're going further, with a Figma agent available directly on the canvas and in the left rail.
That's why we built the Figma agent. Our goal was to create an agent fluent in Figma and native to the way teams work. That meant making Figma itself legible to a model in ways that aren't possible with third-party toolsβwith deep context on your components, tokens, standards, and best practices.
How do Figma's agent and MCP server work together?
Use Figma's agent when you're working on the canvasβit's built in and has additional context about your design system. Use the MCP server and use_figma to pull code onto the canvas or push designs back to code.
The result is a dedicated design agent that works beside you on the canvas, inside the same file as your teamβas a true collaborator. It's fine-tuned for editing Figma files, so outputs are tailored to your design context and built for direct manipulation so you can stay in control. Unlike the MCP server, the agent lives directly on the canvasβno separate setup or context switching, so you can easily:
Start a prompt from any design layer
Parallel prompt to play out multiple ideas at once
Make edits and iterate while the agent iterates, too
From exploring new directions to making bulk edits and implementing feedback, here's how Figma's agent will fit into your design workflow today.
The best designs rarely come from the first ideaβor the first prompt. Exploring directions, comparing approaches, and iterating is already core to how designers work. Our agent will help you cover more ground in less time.
How do Figma Make and the Figma agent fit into your workflow?
You can start in Figma DesignΒ withΒ our agent to generate design layers to clarify intent across flows, states, copy, and structure. Then, send it to Make to generate code layers to clarify behavior. Then, embed it back in Figma Design.
Or, start in Figma Make, copy frames to Figma Design, iterate with our agent, and send it back to Make.
You can go wide: Quickly generate distinct stylistic approaches to the same problem and explore many directions at once. Compare multiple checkout flows optimized for different business goals. Ask the agent for three different information architectures.
Or go deep: Pick a direction and ask the agent to iterate, compare implementations, and rethink existing designs while staying aligned with your design system. The agent uses your most frequently and recently used components as a starting point, but you can also guide the outputs by choosing a specific library and @ mentioning specific tokens, variables, and components to steer it toward exactly what you need. Think of it as a key command for your design system.
As AI makes it easier to generate designs, the risk is shipping something average. Our agent is purpose-built to help you explore more directions, so you can choose the right oneβthen once you have, going hands-on is often faster, more natural, and token efficient than prompting your way to the ideal output. You can also build simultaneously with your team in the same workspace, without losing context.
Design exploration prompt: Give me 3 style options for this design: one that's organic, one modern, and one retro.
Design generation prompt: Make a horizontally scrolling image carousel for a mobile app. Populate the images with placeholder outdoor/sports imagery. Give me multiple options with different title treatments with text on top of the image and below to see what works best.
With the Figma MCP server work can move between code and Figma without losing fidelity. Start in code, bring it into Figma with our code-to-canvas capability to iterate or apply your design system, then send it back via the Figma MCP serverβand everything stays in sync. Our agent fits into that handoff, helping maintain momentum as work moves between tools.
When Figma's agent works alongside you on the canvas, you don't just work fasterβyou can shift easily between AI assistance and hands-on manipulation. This makes it easy to work through tedious tasks that still require context and precision. Wouldn't it be great if you didn't have to manually rename variables for consistency, swap the same component across many different screens, repeat a padding change across an entire flow, or populate a large number of frames with realistic content? That's exactly what the agent is for. It handles the work that usually slows you down by automating bulk edits, applying design systems, and filling designs with realistic content at scale.
Chat with Figma to update typography across a file, replace lorem ipsum copy and imagery across a grid, set all chip components to their active state, or convert your screens to dark mode without manually adjusting fills and contrast. And if you maintain a design system, the agent is a particularly powerful ally. Use it to bulk-update descriptions, tags, and use cases across libraries, standardize naming conventions, document components with all their states and variants, and steer quality by sharing examples of your team's work.
Design system consistency parallel prompts: Set all @ Tile / Callout components to their highlight states; find all text larger than size 10 and change font to Inter; use a cleaner typefaceβββββο»Ώβο»Ώββββββο»Ώο»Ώβο»Ώβββββββββο»Ώββββββο»Ώββββββο»Ώβββββββο»Ώβο»Ώββββββο»Ώββββββο»Ώβββο»Ώββββο»Ώβββββββο»Ώο»Ώββββββο»Ώββββββββββο»Ώβββββββββββββββο»Ώβββββββββββο»Ώβββο»Ώβββο»Ώβββο»Ώβο»Ώβο»Ώββββο»Ώο»Ώββο»Ώο»Ώβββββββββββο»Ώββο»Ώβββββββο»Ώββο»Ώβο»Ώββββββο»Ώββββββο»Ώβββο»Ώββββο»Ώββο»Ώβο»Ώβο»Ώβββο»Ώββββββββββββο»Ώο»Ώββο»Ώο»Ώββββββο»Ώββο»Ώββββββββο»Ώββο»Ώββββο»Ώο»Ώβββββββββββββο»Ώββββο»Ώο»Ώββο»Ώβββο»Ώο»Ώβββββββββο»Ώο»Ώββο»Ώβββο»Ώβββββββο»Ώβο»Ώββββββο»Ώββο»Ώβββββββο»Ώββββββββο»Ώο»Ώββο»Ώββο»Ώβο»Ώβββββββββο»Ώο»Ώββββββββββββββο»Ώβο»Ώβο»Ώβββο»Ώβββο»Ώββββββββο»Ώββο»Ώβββο»Ώβο»Ώβο»Ώββββββββο»Ώββο»Ώβββο»Ώβββο»Ώβββο»Ώββββο»Ώββο»Ώβββο»Ώβββββο»Ώβο»Ώββββο»Ώββο»Ώβββο»Ώβο»Ώβο»Ώβο»Ώβο»Ώβββο»Ώβββο»Ώβββο»Ώβββββο»Ώβο»Ώβββββο»Ώβββββββββο»Ώβο»Ώβο»Ώβββο»Ώβββο»Ώβββββββο»Ώο»Ώββββββββββββο»Ώββο»Ώβββββββββββο»Ώβββο»Ώο»Ώβββο»Ώββββββο»Ώο»Ώβο»Ώβο»Ώβο»Ώβββο»Ώβο»Ώβο»Ώβββββββο»Ώβββββββο»Ώο»Ώββββο»Ώββο»Ώο»Ώββο»Ώββο»Ώββββββββο»Ώββο»Ώβββββββο»Ώββββββββο»Ώο»Ώβββο»Ώβββββο»Ώβββββββο»Ώβο»Ώβββββββββο»Ώβο»Ώβββββββββο»Ώβββο»Ώβββο»Ώβββο»Ώβο»Ώβββββββββο»Ώβββο»Ώβο»Ώβο»Ώβο»Ώβο»Ώβββββββο»Ώβββο»Ώβββο»Ώβο»Ώβββββο»Ώβββο»Ώβββββββο»Ώββββββββο»Ώββββββο»Ώβο»Ώβο»Ώβο»Ώβββββο»Ώβββο»Ώβο»Ώβββββο»Ώββββββββο»Ώο»Ώβββο»Ώβββββο»Ώβββββββο»Ώβο»Ώβββββββββο»Ώβο»Ώβββββββββο»Ώβββο»Ώβββο»Ώβο»Ώβο»Ώβββββββο»Ώβββββο»Ώβο»Ώβο»Ώβο»Ώβββο»Ώβο»Ώβββββββββββο»Ώβββββββββο»Ώβββο»Ώβββββββο»Ώββββββββο»Ώββββο»Ώβββββο»Ώβββο»Ώββββββββο»Ώο»Ώββο»Ώββο»Ώο»Ώο»Ώβββββββββο»Ώβο»Ώβββββββββο»Ώββββο»Ώββο»Ώο»Ώββββββο»Ώβββο»Ώβββο»Ώβββο»Ώβο»Ώβββββο»Ώβο»Ώβββββββββο»Ώβββββββββββο»Ώβββββββββββββββββο»Ώββο»Ώβββββββο»Ώββο»Ώβο»Ώββββββο»Ώββββββο»Ώβββο»Ώββββο»Ώββο»Ώβο»Ώβο»Ώβββο»Ώββββββββββββο»Ώο»Ώβββββββββββββο»Ώο»Ώββββββββββββββο»Ώβο»Ώβο»Ώβββο»Ώβββο»Ώββββββββο»Ώββο»Ώβββο»Ώβο»Ώβο»Ώββββββββο»Ώββο»Ώβββο»Ώβββο»Ώβββο»Ώββββο»Ώββο»Ώβββο»Ώβββββο»Ώβο»Ώββββο»Ώββο»Ώβββο»Ώβο»Ώβο»Ώβο»Ώβο»Ώβββο»Ώβββο»Ώβββο»Ώβββββο»Ώβο»Ώβββββο»Ώβββββββββββββο»Ώβββο»Ώβββο»Ώβββββββο»Ώο»Ώββββββββββββο»Ώββο»Ώβββββββββββο»Ώβββο»Ώο»Ώβββο»Ώββββββο»Ώο»Ώβο»Ώβο»Ώβο»Ώβββββββο»Ώβββββββο»Ώβββββββο»Ώο»Ώββββο»Ώββο»Ώο»Ώββο»Ώββο»Ώββββββββο»Ώββο»Ώβββββββο»Ώββββββββο»Ώο»Ώβββο»Ώβββββο»Ώβββββββο»Ώβο»Ώβββββββββο»Ώβο»Ώβββββββββο»Ώβββο»Ώβββο»Ώβββο»Ώβο»Ώβββββββββο»Ώβββο»Ώβο»Ώβο»Ώβο»Ώβο»Ώβββββββο»Ώβββο»Ώβββο»Ώβο»Ώβββββο»Ώβββο»Ώβββββββο»Ώββββββββο»Ώββββββο»Ώβο»Ώβο»Ώβο»Ώβββββο»Ώβββο»Ώβο»Ώβββββο»Ώββββββββο»Ώο»Ώβββο»Ώβββββο»Ώβββββββο»Ώβο»Ώβββββββββο»Ώβο»Ώβββββββββο»Ώβββο»Ώβββο»Ώβο»Ώβο»Ώβββββββο»Ώβββββο»Ώβο»Ώβο»Ώβο»Ώβββο»Ώβο»Ώβββββββββββο»Ώβββββββββο»Ώβββο»Ώβββββββο»Ώββββββββο»Ώββββο»Ώβββββο»Ώβββο»Ώββββββββο»Ώο»Ώββο»Ώββββββο»Ώο»Ώβ.
Design system documentation prompt: Create documentation for my components, including all states and variants. Add contextual examples illustrating how these components should be usedβββββο»Ώβο»Ώββββββο»Ώο»Ώβο»Ώβββββββββο»Ώββββββο»Ώββββββο»Ώβββββββο»Ώβο»Ώββββββο»Ώββββββο»Ώβββο»Ώββββο»Ώβββββββο»Ώο»Ώββββββο»Ώββββββββββο»Ώβββββββββββββββο»Ώβββββββββββο»Ώβββο»Ώβββο»Ώβββο»Ώβο»Ώβο»Ώββββο»Ώο»Ώββο»Ώο»Ώβββββββββββο»Ώββο»Ώβββββββο»Ώββο»Ώβο»Ώββββββο»Ώββββββο»Ώβββο»Ώββββο»Ώββο»Ώβο»Ώβο»Ώβββο»Ώββββββββββββο»Ώο»Ώββο»Ώο»Ώββββββο»Ώββο»Ώββββββββο»Ώββο»Ώββββο»Ώο»Ώβββββββββββββο»Ώββββο»Ώο»Ώββο»Ώβββο»Ώο»Ώβββββββββο»Ώο»Ώββο»Ώβββο»Ώβββββββο»Ώβο»Ώββββββο»Ώββο»Ώβββββββο»Ώββββββββο»Ώο»Ώββο»Ώββο»Ώβο»Ώβββββββββο»Ώο»Ώββββββββββββββο»Ώβο»Ώβο»Ώβββο»Ώβββο»Ώββββββββο»Ώββο»Ώβββο»Ώβο»Ώβο»Ώββββββββο»Ώββο»Ώβββο»Ώβββο»Ώβββο»Ώββββο»Ώββο»Ώβββο»Ώβββββο»Ώβο»Ώββββο»Ώββο»Ώβββο»Ώβο»Ώβο»Ώβο»Ώβο»Ώβββο»Ώβββο»Ώβββο»Ώβββββο»Ώβο»Ώβββββο»Ώβββββββββο»Ώβο»Ώβο»Ώβββο»Ώβββο»Ώβββββββο»Ώο»Ώββββββββββββο»Ώββο»Ώβββββββββββο»Ώβββο»Ώο»Ώβββο»Ώββββββο»Ώο»Ώβο»Ώβο»Ώβο»Ώβββο»Ώβο»Ώβο»Ώβββββββο»Ώβββββββο»Ώο»Ώββββο»Ώββο»Ώο»Ώββο»Ώββο»Ώββββββββο»Ώββο»Ώβββββββο»Ώββββββββο»Ώο»Ώβββο»Ώβββββο»Ώβββββββο»Ώβο»Ώβββββββββο»Ώβο»Ώβββββββββο»Ώβββο»Ώβββο»Ώβββο»Ώβο»Ώβββββββββο»Ώβββο»Ώβο»Ώβο»Ώβο»Ώβο»Ώβββββββο»Ώβββο»Ώβββο»Ώβο»Ώβββββο»Ώβββο»Ώβββββββο»Ώββββββββο»Ώββββββο»Ώβο»Ώβο»Ώβο»Ώβββββο»Ώβββο»Ώβο»Ώβββββο»Ώββββββββο»Ώο»Ώβββο»Ώβββββο»Ώβββββββο»Ώβο»Ώβββββββββο»Ώβο»Ώβββββββββο»Ώβββο»Ώβββο»Ώβο»Ώβο»Ώβββο»Ώβββο»Ώβο»Ώβο»Ώβββο»Ώβββο»Ώβββββββββββββββββο»Ώβο»Ώβββββββο»Ώβββο»Ώβββββββο»Ώββββββββο»Ώββββο»Ώβββββο»Ώβββο»Ώββββββββο»Ώο»Ώββο»Ώββο»Ώο»Ώο»Ώβββββββββο»Ώβο»Ώβββββββββο»Ώββββο»Ώββο»Ώο»Ώββββββο»Ώβββο»Ώβββο»Ώβββο»Ώβο»Ώβββββο»Ώβο»Ώβββββββββο»Ώβββββββββββο»Ώβββββββββββββββββο»Ώββο»Ώβββββββο»Ώββο»Ώβο»Ώββββββο»Ώββββββο»Ώβββο»Ώββββο»Ώββο»Ώβο»Ώβο»Ώβββο»Ώββββββββββββο»Ώο»Ώβββββββββββββο»Ώο»Ώββββββββββββββο»Ώβο»Ώβο»Ώβββο»Ώβββο»Ώββββββββο»Ώββο»Ώβββο»Ώβο»Ώβο»Ώββββββββο»Ώββο»Ώβββο»Ώβββο»Ώβββο»Ώββββο»Ώββο»Ώβββο»Ώβββββο»Ώβο»Ώββββο»Ώββο»Ώβββο»Ώβο»Ώβο»Ώβο»Ώβο»Ώβββο»Ώβββο»Ώβββο»Ώβββββο»Ώβο»Ώβββββο»Ώβββββββββββββο»Ώβββο»Ώβββο»Ώβββββββο»Ώο»Ώββββββββββββο»Ώββο»Ώβββββββββββο»Ώβββο»Ώο»Ώβββο»Ώββββββο»Ώο»Ώβο»Ώβο»Ώβο»Ώβββββββο»Ώβββββββο»Ώβββββββο»Ώο»Ώββββο»Ώββο»Ώο»Ώββο»Ώββο»Ώββββββββο»Ώββο»Ώβββββββο»Ώββββββββο»Ώο»Ώβββο»Ώβββββο»Ώβββββββο»Ώβο»Ώβββββββββο»Ώβο»Ώβββββββββο»Ώβββο»Ώβββο»Ώβββο»Ώβο»Ώβββββββββο»Ώβββο»Ώβο»Ώβο»Ώβο»Ώβο»Ώβββββββο»Ώβββο»Ώβββο»Ώβο»Ώβββββο»Ώβββο»Ώβββββββο»Ώββββββββο»Ώββββββο»Ώβο»Ώβο»Ώβο»Ώβββββο»Ώβββο»Ώβο»Ώβββββο»Ώββββββββο»Ώο»Ώβββο»Ώβββββο»Ώβββββββο»Ώβο»Ώβββββββββο»Ώβο»Ώβββββββββο»Ώβββο»Ώβββο»Ώβο»Ώβο»Ώβββο»Ώβββο»Ώβο»Ώβο»Ώβββο»Ώβββο»Ώβββββββββββββββββο»Ώβο»Ώβββββββο»Ώβββο»Ώβββββββο»Ώββββββββο»Ώββββο»Ώβββββο»Ώβββο»Ώββββββββο»Ώο»Ώββο»Ώββββββο»Ώο»Ώβ.
Design work accumulates feedback that's often scattered across comments and files: from crit notes and stakeholder reactions to open questions. Our agent helps you put that feedback to work.
Because your whole team works in the same file, the agent already has that contextβso asking it to work through feedback feels less like briefing a new collaborator and more like thinking out loud with someone whoβs already in the room.
Ask Figma to summarize feedback, identify themes, and turn input into next steps. You can also use it to pressure-test designs from different points of view. Wondering how a VP focused on revenue will react? Our agent can model that perspective. Working through a long comment thread? The agent can distill the back-and-forth into an actionable plan. Need to think out loud before a crit? Use our agent to sharpen your direction before sharing it with your team.
Organize and apply feedback prompt: Sort these comments by theme. Create a new rev of my design that incorporates the feedback for this profile.
Figma's agent is embedded where the work already happens. There's no toggle tax, no context switching, no learning curve. You stay in Figma and your team stays in the loop. We built this with one goal: to help you work faster without compromising on quality and craft.
In the coming months, we'll focus on improving support for design systems, refining the user experience, expanding the agent's ability to search across your files, and adding ways to customize. Weβre just getting started.
The Figma agent is rolling out gradually in beta over the coming weeks. During beta, the agent won't consume credits. AI credits will apply at general availability. Sign up here to request early access. If youβre eligible and selected for the beta, weβll send you an email once youβre in. Joining the list does not guarantee access. The agent will be available for Full seat users on Professional, Organization, and Enterprise plans. Collab and Dev seats can use the agent in drafts. Starter, Education, and Government plans are not included. Learn more in our help center.
Rodrigo is a product manager at Figma, working on agents. Before Figma, Rodrigo led AI product management at Asana.