Is SolidCraft supposed to run in the browser or do I need to build it locally?

Oct 22, 2026 Last reply: 10 hours ago 6 Replies

First reply

MesaLynx994

There's no hosted version as far as I can tell. The repo only ships build instructions, and the person behind it seems focused on the desktop app right now. What I did on my old laptop was compile the webassembly target myself and serve the output folder with any static file…

Read the full reply

I keep seeing people talk about SolidCraft being usable right in the browser, but when I go to the repo I can only find instructions for building it from source with Rust. Do I need to compile the whole thing myself or is there a hosted web version somewhere I'm just not finding?



I got the desktop build going on my machine after some fiddling with the toolchain, and honestly it feels snappy for somethig so new. But my laptop is old and weak, so I'd love to run the browser version there instead if it exists.



Also, when I open a new file I get this timeline thing on the side like Fusion has, but half the time my sketch features refuse to reorder and the whole model just errors out. Is that a known thing or am I doing something dumb? Feels wierd to be this early to a project where the first commits are barely two weeks old, so maybe I'm just beta testing without meaning to.


There's no hosted version as far as I can tell. The repo only ships build instructions, and the person behind it seems focused on the desktop app right now. What I did on my old laptop was compile the webassembly target myself and serve the output folder with any static file server. Since it all runs client side, once you've got the build artifacts you can literally copy them onto a USB stick and open them from the laptop, or toss them on free static hosting if you have any. It's not the same as a maintained hosted instance, but it means you only need the toolchain on one machine. One caveat: the web build felt noticeably slower than the desktop one on my ancient ThinkPad, so test it before you assume it'll be lighter. Heavy geometry seems to choke the browser build more than the native one.

On the timeline reordering, you're not doing anything dumb, that's a known rough edge. When I try to drag a sketch feature past something that depends on it, the whole chain errors out and sometimes the errors don't clear even after I drag it back. What's worked for me is reordering in small steps, one slot at a time, and checking wether the feature references anything downstream before moving it. If it does error out, undo to the last good state rather than trying to fix it in place, because the error state tends to stick.

And yeah, you basicly are beta testing without meaning to. With a codebase this young the file format itself might change, so I'd keep exports of anything you care about and don't treat the native project files as safe storage. If you hit the reorder bug consistently, log it in the tracker with the exact feature order, stuff like that gets fixed fast when the steps to reproduce are clear.

Glad I'm not the only one hitting the reorder thing, thought I was losing it. My build actualy went smoother than I expected, took maybe 20 minutes once I stopped fighting an outdated compiler. Curious if anyone knows if the dev posts a roadmap anywhere, or is it just commit and pray? Feels like the kind of project that could go quiet overnight and leave a bunch of us with files we can't open anywhere else.

Seconding the static hosting trick, it worked well for me too. I served the wasm build with python http.server on my desktop and just opened the laptop browser to my local IP, so I never had to touch the toolchain twice. Only snag was the laptop needed a browser update first, the older chrome I had on there would load the page but the canvas stayed blank. On the timeline bug, mine got so bad mid-session once that I had to close and reopen the file before the errors cleared, so a restart is worth trying if undo stops working.

The feature order thing lines up with what I hit too, but mine got worse when sketches were referenced across the timeline instead of inside one body. I had a job last week with 14 features on a mounting bracket and moving even one sketch plane froze the whole chain til I rebuilt from scratch. Now I sketch everything on the origin planes first and keep references to those only, seems to dodge most of it. Curious whether the dev is tracking this in the issue list or if we should all pile our repro steps into one thread.

Been bitten by young parametric tools before, back when we rebuilt half a model because the history tree choked on a fillet. My habit these days is to treat any feature tree under two years old as disposable. I sketch, I export a mesh every hour or so, and I never count on the timeline surviving a session.

One thing worth trying on the reorder bug: the repo is on git, so you can bisect. Check out an older commit, rebuild, and see if the drag works there. I did that on a project years ago and found the exact commit that broke it, wich makes for a much better bug report than a vague 'it errors out'.

Update: the wasm build plus python http.server trick worked, my laptop now runs the browser version fine off my desktop. Still stuck on the reorder thing though, my sketches keep breaking the timeline the same way. For the person with the 14 feature bracket, did rebuilding actually fix it for good or did it come back?

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required