I work on Arabic and multilingual product interfaces, and I kept running into the same problem: converting an LTR design to RTL was rarely just a visual flip.
Horizontal Auto Layout with Wrap is a good example. Reversing the child order can align the first row, but it does not reliably fix the direction of every wrapped row. It can also make responsive components harder to maintain.
I built RTLAuto to handle this workflow inside Figma.
Who RTLAuto is for:
- Product designers maintaining Arabic and English versions
- Design-system teams working with reusable components and variants
- UX and UI agencies delivering multilingual products
- Product teams building for Arabic, Hebrew, Persian, or Urdu markets
- Anyone spending too much time rebuilding RTL screens manually
In version 1.3.1, the current conversion path has dedicated support for:
- Horizontal Auto Layout with Wrap
- Nested Auto Layout
- Reusable components and direction variants
- Responsive wrapping without absolute positioning
- Directional icons while protecting semantic icons
- Mixed Arabic and English content
- Text Styles, text properties, and Variables
- Non-destructive LTR-to-RTL copies
For Wrap specifically, RTLAuto keeps the logical child order stable, mirrors the horizontal row flow for RTL, and keeps managed children inside Auto Layout so the result remains responsive. For reusable component variants, the original component structure is preserved and the RTL flow is isolated in the generated direction variant.
I verified the current 1.3.1 build before posting this: the Wrap transformer tests, component variant tests, and semantic icon tests all pass.
The product image and official demo video are attached directly to this topic.
Try RTLAuto on Figma:
https://www.figma.com/community/plugin/1623814570395556605/rtlauto
For transparency, I am the maker of RTLAuto. I would especially value feedback from designers working with real multilingual products. If you have a difficult Wrap, component, or design-system case, please share what breaks today. I want to test the plugin against production patterns, not only ideal examples.
