# New WebGL rendering engine

**URL:** <https://community.penpot.app/t/new-webgl-rendering-engine/10731>\
**Category:** Feature requests\
**Created:** [July 18, 2026, 4:09am UTC](https://community.penpot.app/t/new-webgl-rendering-engine/10731 "2026-07-18T04:09:35Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![mkeida](https://avatars.discourse-cdn.com/v4/letter/m/5e9695/32.png) [@mkeida](https://community.penpot.app/u/mkeida)\
**Post date:** [July 18, 2026, 4:09am UTC](https://community.penpot.app/t/new-webgl-rendering-engine/10731/1 "2026-07-18T04:09:35Z")

</div>

Hi everyone,

I wanted to share some honest feedback after a significant investment on our end.

We spent roughly **72 hours of human work** migrating **all of our company’s designs** from Figma and Adobe XD into Penpot — around **56 files** in total. We were genuinely optimistic about this move, especially given the promises around the new rendering engine. We expected it to be the safe, reliable choice going forward for our team.

Unfortunately, the vast majority of the migrated files turned out to be **essentially unusable** due to serious performance problems. Ironically, the new rendering engine performs **worse than the old Canvas-based solution** in our real-world use case.

Specifically, working with boards/artboards across roughly **9 screens** in a single file grinds Penpot to a halt — panning, zooming, and selecting objects become extremely laggy to the point of being unworkable for actual production work. I’m attaching a video that demonstrates this directly.

We really wanted this migration to succeed — the tool has a lot going for it philosophically (open source, self-hostable, no vendor lock-in), but performance at this level is a dealbreaker for teams with non-trivial file complexity.

At this point, I have to be blunt: **I strongly discourage anyone from relying on Penpot for real production work.** It’s a nice demo and a cool experiment, but honestly, I wouldn’t waste any more time on it. After 72 hours of migration work across 56 files, ending up with something slower than our old Canvas-based setup is not a good look, and it’s hard to trust the roadmap after this.

Video: [Private Seafile](https://200.solutions/seafile/f/b59a75caf27a4d92afd6/)

---

<div class="post-metadata">

**Author:** ![a-ng-d](https://dub1.discourse-cdn.com/flex017/user_avatar/community.penpot.app/a-ng-d/32/11021_2.png) [@a-ng-d](https://community.penpot.app/u/a-ng-d)\
**Post date:** [August 14, 2026, 5:02pm UTC](https://community.penpot.app/t/new-webgl-rendering-engine/10731/2 "2026-08-14T17:02:18Z")

</div>

Hi,  
I understand your frustration, but what haven’t you migrated the largest files first in order to check some acceptance and unnegotiable criteria before?

---

<div class="post-metadata">

**Author:** ![elenatorro](https://dub1.discourse-cdn.com/flex017/user_avatar/community.penpot.app/elenatorro/32/7407_2.png) [@elenatorro](https://community.penpot.app/u/elenatorro)\
**Post date:** [August 17, 2026, 7:33am UTC](https://community.penpot.app/t/new-webgl-rendering-engine/10731/3 "2026-08-17T07:33:19Z")

</div>

Hi, Marek!

We keep working on performance actively by improving the rendering algorithm. Sometimes we discover there’s a corner case we haven’t covered that might impact performance negatively, like a specific file structure. If you could give us a file example where you can reproduce the regression, that will help us to figure out where the problem is and fix it. This kind of feedback is super helpful for us!

Thanks!