Share Your Thoughts

I've been using the Figma MCP server with Claude Code to build a design-to-code workflow, and I've hit what feels like a significant gap in the `generate_figma_design` (html-to-design) capture tool. I'd love to see this addressed and wanted to share the use case in detail. The current behaviorWhen generate_figma_design captures a rendered web page, it converts everything to raw Figma geometry — frames, rectangles, and text nodes. Even if the page is built entirely from your published design system components (Button, Tag, Checkbox, etc.), and even if those components have Code Connect mappings configured, the capture output has no awareness of them. You get a pixel-faithful snapshot, but it's "dead" — no component instances, no variable bindings, no variant properties. The gap this createsCode Connect already establishes a bidirectional relationship between Figma components and code components. The metadata is there. But generate_figma_design doesn't use it. Here's a concrete example o

Greetings, I’ve been using an AI agent to develop a prompt to review figma concepts to cover the following: UX heuristic using NNg as the foundation, UI errors, and Grammatical errors. The prompt needs to be copied from the Ai tool and implemented in Figma Make. Figma Make has to recreate the reviewed design exactly before this prompt is implemented to review the design. This is a lot of steps to review a designs. Will Figma introduce a a feature to fully review designs on an artboard covering those items? UX errors, UI errors, and grammatical errors?

Request: More access through the Figma MCP when inspecting the node in `dev` mode.Reason: I’d assume (correct me if wrong) this would decrease token usage because of the extra coding data (e.g., setting css styles to `Android`) that Figma already has available.Open to being corrected by the Figma team!

Problem As a front-end programmer I don’t quite grasp the current concept of so called “Local Variables”. For starters, they’re not local, but global. And the thing with global vars is - they make a big mess, like always. For this reason any programming language considers global vars an anti-pattern. But currently, the only way to make a Figma component “stateful” is to keep its “state” in the global namespace. I use quotes for a reason: there is no real state at all, because in Figma there is no an entity like “state of a component”. This all makes flows unmaintainable and prevents us from transferring components between files and projects, because they get transferred without their “state”. So basically you can’t easily reuse components logic. It also doesn’t provide any level of isolation. For example, if I have more than one instance of the same component, how would I distinguish their variables/states? Solution proposal Since Figma’s model is a tree-like, and frames serve as the

Mural’s tables auto adjust based on items within a cell. If you add another sticky note, the cell expands to accommodate the new sticky. Not having this in Figjam is killing me. 

Currently, when we have many files open in Figma (which is not uncommon in a work context), we tend to lose visibility and access to some of them when trying to switch tabs.Intuitively, I try to scroll horizontally through the open file tabs with no success, which makes navigation quite difficult. As a result, I either have to drag the entire Figma window to the side to expand the view or feel forced to close files just to access others.Is there any plan to improve this usability?Thank you.

It would be a big time saver to preserve the style and formatting of a text layer, when copying and pasting it from one file to another file. Right now the paste operation loses all formatting and I have to re-format the text all over again.

When working with multiple shapes, you currently cannot edit the anchor points of multiple shapes at once whether they are unionized or not. At my studio, we are working on incredibly complex UI for vehicles, and products, and this feature would be much appreciated, and would greatly increase the speed of our workflow.

I would love edit multiple images at the same time and not have to wait till one is done! Very time consuming to have to wait when you have to do 6 or 10.

Searched and didn’t see any mention of this. If you have a file with a blue background that you wanna send to someone to test it can be rough. Custom colors would be cool, but at least being able to pick from like 2-4 would suffice. Thanks!

I use several whiteboards with my students.  We always use the marker option.  The marker option has thin and thick for the stroke mark.  For some students the thin/thick option are think on others it is thin. Looking at the pictures below, how can I get all of my boards to allows for the board 2 marker strokes.  I need thinner options. the screenshots from the different boards prove this is possible.  

This is a reference to a topic that was closed without resolution: Reorder objects in an autolayout component instance. Allow for order overrides Share an idea Reorder objects in an autolayout component instance. Allow for order overrides. Allow users to reorder (drag or arrow) objects inside a component instance that has auto-layout on it. Especially useful for components that have repeated elements, such as a dropdown menu or a group of form inputs, navigation. This would be much quicker than manually changing each of the elements in-between where the current element is and where the new position is. Or for users that don’t have patience and break… I can’t express how frustrating it is when a designer has to detach one of your components simply to add a tab in-between existing tabs or a dropdown option above existing ones. When I build a component I shouldn’t have to create 10 hidden layers as placeholders to make sure that the component works for a

Simple one - slots should keep their pink color even when not selecting or dragging. 

I do not want my file to be copied. I just want the file that I share to export that can then also be exported. I need more options. If I create a export the is possible.

I’ve just seen that I can’t get back to the instance I was just working on by clicking on the “Return to Instance” button.

In Figjam - When adding a new node from an existing one, Holding down SHIFT (or any reasonable hotkey) should reverse the arrow to point in to the current node.This will reduce/ eliminate the work to reverse the arrows while making diagrams.  

Hi,We have thousands of icons as components in our library that should not be linked via Code Connect.It would be great to have a way to multi-select them and apply a batch action, such as “Skip connection”.At the moment, I would need to click on each icon, select “Skip connection” on the next screen, then go back to the list… Something like this, used in combination with filtering by name would be great!Thanks

How can I view Figma prototypes on my mobile, or is that currently only for web desktop?When I copy the share link and open in on my mobile, it opens up in a very small view and completely unusable on phone.

SummaryExpose presenter notes as a read/write property on SlideNode in the Plugin API, so plugins and MCP servers can programmatically get and set presenter notes for Figma Slides.Proposed APIinterface SlideNode { // ... existing properties ... presenterNotes: string // read/write access to the slide's presenter notes}Use CaseI'm using the figma-slides-mcp server to create and edit Figma Slides presentations programmatically from AI coding agents (Claude Code, Cursor, etc.). The MCP server can already create slides, set text, clone templates, reorder slides, and take screenshots, all through the Plugin API.The one important gap is presenter notes. When generating a slide deck programmatically, the presenter notes (speaker script, talking points, timing cues) are often as important as the content on the slides (I use the presenter notes as teleprompter scripts in my online classes when I’m giving recorded lectures). Right now there is no way to set them. SlideNode has no presenterNot

Some quick context: I recently created a basic Figma slides template. After using the template & theme to create an actual presentation, I identified some new slides that I would like to include into the original template. When copy pasting the slides that used the template into the original template file, it created a duplicate template with the same name for the slides that were pasted in, even though they used the original theme of the template. The main issue I have here is that when swapping the template to the original, it randomly shuffles the colors. So I have to adjust the entire slide. I feel like updating templates over time based on real usage is pretty common, and the current behavior makes it rather awkward. Thanks!

I want to pay for the design of the Figma Make AI, but I don't want to pay for the subscription, as I'm not a web designer and I don't think I'll be using Figma for any other projects anytime soon.That's why I need to know if it's possible to pay only for the finished product and not the monthly/annual fee.Thanks in advance,

I'm using a chat agent to prototype my designs. But currently in Figma mobile app, it only allow me to view and comment on my canvas. I cannot select the frame and share the frame link to my chat agent as a reference for vibe coding. The workaround of request desktop app from Browser crashes all the time. Please enable the feature to be compatible with the recent AI trends.

Why figma do not copy comments with page or element. when i copy some object that have client comment or feedback.

 it’s been months since this issue above was reported and asked, and i saw a solution here  however it’s too tedious and it’s not even responsive. i also have to create it on each page. When would devs add overlay feature on figma sites? overlay is such a basic animation on all sites. i only have free tier for school is the overlay feature locked behind a paywall?

Switching to the Enterprise license and would like the ability to assign a user to multiple workspaces. I see there is an option for folks to be invited to teams in different workspaces, but it would be more efficient if we can easily add and see how many workspaces folks are assigned to. In our organization we do a lot of cross collaboration, across the workspaces.