Suggest a Feature

As a regular Figma full seat user, I couldn’t use all the credits during the holidays and when I fell sick for a week. Can we please have credits rollover so that we get to use what we paid for?Many Thanks!

We've been excited to integrate Figma MCP into our workflow to connect and update our design system via claude code, but ran into a significant wall: the remote MCP server at mcp.figma.com only works with a personal Anthropic Pro or similar account.  We're required to use Claude Code via AWS Bedrock here, which is how many enterprise teams access Claude.  Figma MCP does not support Bedrock or other enterprise auth paths.We spent a couple of hours troubleshooting and found that those of us using the AWS bedrock keys can't use the MCP server setup that is documented. Instead, they have to:Install Node.js manually Create a background service (launchd plist) to run the MCP server locally Manually configure Claude Code to point to a local HTTP server on port 3333 Debug port conflicts and token issues that the remote setup handles automaticallyWe don't have it fully working yet, but this is already a meaningful barrier for the tea

Hi! I'm the developer of Vibma, an MCP plugin that was recently rejected from the community because it overlaps with Figma's official MCP write access. I understand the rationale — two plugins shouldn't compete on the same "bring your own AI into Figma" surface, and I agree with that decision.But I want to propose a path forward, because I think the rejection conflated two different kinds of value:1. AI access — letting external models (Claude, GPT, etc.) read and write Figma files. This is what the official MCP already solves, and I agree Figma should own it.2. The harness — the structural guardrails that sit between any AI and the file. In Vibma, that's things like:Enforcing auto-layout instead of absolute positioning Rejecting hardcoded colors and nudging toward variables Linting for unbound tokens, missing text styles, component misuse Batching and validating mutations so

To reduce component weight, my design system team has been exploring using variable modes to toggle between different “variants” of a component, rather than just using variants. The benefit of this is that for these “style” specific properties, where just styles are changed like fill and stroke, using modes significantly reduces the weight of a component in the system. In this example I use a Badge component, and variables to change between the colors of the badge.However, my component code still has a “color” prop. Because I don’t have a color prop on my Figma component, but rather a “BadgeColor” mode that controls this property, is there a way in which CodeConnect can still map this to the correct code property? Currently it doesn’t seem like CodeConnect can read anything that is not a property. Realistically all I need is the CodeConnect to be able to read the string value of the mode, and map that to the actual code prop.

At the moment if I open a project with one account, and want to edit/comment on another account, I have to go to Home/recently viewed and switch accounts. It would be great if I could ‘switch accounts’ within the project itself.

Hi Figma team,I'm using the Figma remote MCP server (https://mcp.figma.com/mcp) with Kiro (by Amazon). The connection works fine — I can successfully use read-only tools like get_design_context, get_metadata, get_screenshot, etc.However, I noticed that the use_figma tool (write to canvas) is not available when connecting from Kiro. According to the official documentation table at Guide to the Figma MCP server, Kiro's "Write to canvas" column is empty, while VS Code, Cursor, and Claude Code all have it enabled.I confirmed this by connecting the same MCP server from VS Code — use_figma appeared immediately with the full toolset. So the limitation is clearly on the server side based on client identification.Request:Please consider enabling use_figma and search_design_system tools for Kiro, the same way they are enabled for VS Code and Cursor. Kiro already supports Streamable HTTP MCP connections and handles tool approvals properly.Thank you for

I have had a need on multiple occasions to view the release/publish history of a library. I typically include detailed notes when I hit the publish button, however I am unable to find it afterwards for record sake

We raised it in the past:  Hi,We are integrating with Figma as part of our SSPM solution and need access to an API for organization settings and accounts, similar to the one used by the UI. Unfortunately, I couldn’t find documentation for this API. Is it possible to use a PAT (Personal Access Token) of an admin for this purpose? If not, do you have any plans to support this functionality?Thanks! Can you please give us some sort of roadmap for it?

Please can we get percentages for variables? We can use percentages manually, but we can’t create a percentage as a variable. This would speed up my design to dev process tenfold, and considering we can already use percentages in the design, why disallow it in variables?An example is I want store values for line height and letter spacing as percentages which behave responsively/dynamically, so when creating type variables I don’t have to do manual calculations, store 10+ letter spacing/line height values based on which font size they will be used with and hope I never change the font size variable in the future.

Hello, I’m trying to work out if it’s possible to change text colour within a table cell on FigJam. If I choose a change of colour it is only allowing me to change the background colour. I’ve tried hacking it but to no prevail, any ideas? Thanks Please search for existing topics before posting! Press 🔍 at the upper right to sea

Reopening a thread for interest in a presentation mode in FigJam (here’s the original thread that’s now closed). Of course someone can follow a specific user as they zoom and scroll through a board, but the viewer experience–especially for those unfamiliar with FigJam–could be improved if a specific presentation mode existed. Thinking similar to Miro’s, which would mean also needed to add “frame” elements to be able to have it function like a slide deck.

Fractional units and hugs are definitely a welcome addition to Grids, but what’s still missing is the ability to assign both the number of columns/rows and colspan/rowspan to variables.We’ve defined our website breakpoints as modes so we can build components that automatically adapt to the viewport (for example, a 50/50 layout that stacks on mobile). Right now, the only way to achieve this is by storing fixed column-width values for each breakpoint and use them in combination with horizontal auto-layout. It works, but it defeats the whole purpose of Figma’s grid system.Assigning variables directly to the grid would allow every component using that grid to automatically adapt based on its current mode (breakpoint). This would make creating mockups across different breakpoints significantly faster and more efficient.

Hi,I’ve seen this topic mentioned before, but I’d like to propose a more specific feature request for slide layers.It would be very helpful to have selected auto layout options available directly on the slide layer itself, not only within frames placed on top of it. Since slides need to maintain consistent dimensions across the entire presentation, options like “Hug contents” likely wouldn’t apply here and could be excluded.However, enabling flow direction (vertical or horizontal) and potentially grid layout at the slide level would significantly improve layout flexibility. It would allow designers to structure content directly on the slide, without the need to introduce an additional frame solely to activate auto layout across the full slide area.This would streamline workflows and reduce unnecessary nesting, especially for simpler slide compositions.Thanks! 

A great feature to have is text styles not breaking when a designer is using Cmd + B to bold a word or phrase. Currently it unlinks the text style and applies the bold to the underlying font. It would be great if there was a feature where when I use Cmd + B, it swaps to that text style in the library (where the style already exists) and updates in the property panel as well.

I need to mimic a part of the screen expanding upon horizontal scroll of an image gallery 

Hey Figma Team!I have been working on a AI Skill+ Figma Plugin that automates the entire process of Design Token creation in Figma. This product can create enterprise grade variable system in minutes. It supports scoping, code syntax, publishing, proper aliasing, and automatic generation of styles for text, effects and layout grids.I would love it if you could check it out.I made sure this doesn't use a lot of tokens(works on free version of many AI tools, and also with flash models). Repo: https://github.com/Shanmus4/figma-variables-tokens-generatorI would like to share this work with the Figma team, and see if they can integrate this in the product itself. I have attached a small video of how this skill can be used with Claude on my LinkedIn post: 

Hi everyone, I’m trying to export frames to PDF with the raster images as losseless, uncompressed PNGs. I know this is possible because on a very similar project last year, I was able to export the frames that way. As of right now, the PDFs are exporting the raster images with heavy compression/artifacting. I need the export to be PDF as well so the font/text stay vectors. What should I do? Thank you!

Is it possible the prompt result from figma make design wise to be copied in the main design file i’m not seeing any button that allows me to do that 

When a Figma branch is updated from main, components modified on the branch can be replaced wholesale, even when the branch changes are additive or limited to layers or properties that do not overlap with changes from main. In a recent library-maintenance session using Claude and Figma skills, this caused branch-specific edits to be lost multiple times and required rerunning rebuild scripts after each update from main.Request: allow incoming updates from main to be reviewed and applied at a finer granularity than the entire component. Ideally, non-conflicting changes could be preserved automatically, while actual conflicts are surfaced at the property or layer level with a choice to keep the main version or branch version. Even a simpler option to preserve branch changes where they do not conflict would be a meaningful improvement.Use case: I maintain a Figma library that mirrors a shipped React design system. My branch often contains

The Plugin API currently cannot enumerate styles or variables available from libraries linked to the current file. Methods like getLocalTextStylesAsync() and the variable equivalents only return local items, so using linked-library assets requires the plugin to already know the asset keys in advance.This creates a discovery gap for plugins that work with library content dynamically, especially AI-driven workflows that maintain design systems from code. A linked library may already be available in the file, but the plugin cannot inspect its contents unless the keys were prepared ahead of time or the assets are already applied somewhere in the document.Request: add async Plugin API methods to enumerate styles and variables reachable from libraries linked to the current file, without requiring pre-known keys.Use case: I maintain a Figma library that mirrors a shipped React design system and use AI-agent scripts to apply updates as code components c

I’ve been using Buzz to create full page templates, which has great potential. However, in the Edit Content panel, all editable fields for the entire page are displayed at once, which makes it difficult to navigate and manage.My suggestion is to improve this panel by making it context-aware. For example, when a specific section or component is selected on the page, the panel could display only the fields related to that element. This would make editing full pages much easier and more efficient.To give some context, in one of my projects, Buzz shows around 160 text fields along with dozens of media items. That’s a lot of information to manage in a single panel. A filtering mechanism based on the selected element would significantly improve usability.

Our team is part of a larger organization, and all our Figma files are in one Figma Team for which I am a team admin. We have some Default Libraries set for the entire team, but need the additional ability to set Default Libraries for specific project folders. I’m not seeing the settings to do this right now.

I am using Figma from almost 2 years. It’s a great tool. Previously I was using Windows 10 yet recently switched to Linux(Ubuntu). Figma doesn’t have any official release for Linux. There are some Unofficial desktop versions but they are “BUGGY”. I request you guys to release an official Figma desktop app for Linux.

Please make it possible to control which properties of nested components (atoms) can be used in higher components (molecules). What if I use same nested component in different objects and I don’t need some properties in both of them? What if I don’t need some booleans which is part of variants of that atom in molecule? Please fix it, thank you.