The launch graphic says "Ported end-to-end by Fable, an AI agent" and it definitely feels that way. Buggy, glitchy, needs some love and human eyes before it's really usable.
FreeCAD on desktop is much the same in my experience. If LLMs were mainstream a few years ago I would have assumed the UI was 100% AI generated with zero human input besides a oneshot prompt.
When was the last time that you used FreeCAD? Even 18 months ago, I would agree with you that it was a buggy mess, but I tried it a few weeks ago and it was super stable. The UI is still a little confusing, but that just comes with the domain, since AutoCAD and Fusion 360 aren't much better.
Browser can be local. What’s nice about browser based is that browser based programs can run on every device. Though it sounds like this one requires chrome which seems weird to me.
(I made this port) Fwiw I personally had no reason to do this port beyond using it as a benchmark of the agentic capability of Fable, where something of this shape is IMO a way better gauge than those dumb X.com 'I oneshot game with models X/Y/Z this is how it compares'
I published the actual prompts, and you can see quite clearly that vs Opus which is ok at implementing one big feature, Fable was really able to push through a good chunk of the port. That said it definitely didn't one-shot the port, it also didn't figure out a broken docker sbx sandbox by itself, and also later needed some gaslighting into thinking that the port is not really that hard (by any human measure it was quite hard given the scope of code involved.. The nearly 200MB wasm binary is mostly code afaict..). So there are some clear patterns of how the model was trained and also roughly the scope of task visible in those traces. What I see is that it likes prompts that would take an L4/L5 2-4 weeks to do with Cursor ~2 years ago, more needs some direction and deliberate prompting.
Because then it doesn't matter what you're running locally, as long as you've got a supported browser (Chrome, I'm guessing). It means it doesn't have to make a difference if you have a Window 10 desktop or a MacBook Air or a Chromebook. Go to the web page and look at this CAD.
It is fine 'cause it uses the browser you're already using anyway instead of importing a new one for each and every application. Run the backend on a server somewhere and you can get to your application anywhere you can reach that server which - using a self-hosted VPN - can be anywhere you have internet access.
Onshape is free in the browser as long as you are not doing commercial work. It’s a professional system from the founders of Solidworks that competes with all the top CAD tools.
> It’s a professional system from the founders of Solidworks that competes with all the top CAD tools.
Because it's built on Parasolid, the same geometric kernel as everyone else. With ACIS pretty much out the door, almost all the professional CAD packages are just window dressing on the same CG implementation.
Agreed. I’m not OP but for six months I’ve been using Claude to build a from-scratch CAD kernel based on Rust and WASM, MIT licensed.
The actual UI still needs a lot of work, but I’ve been focused on the kernel. Fable has helped a lot though Opus was already making great headway.
I’m an OnShape power user going back about ten years, Solidworks before that. I need a CAD system that absolutely works. There is a lot of work to do still, and it still seems impossible to succeed, but I’ve been very happy with where things have been going with it lately.
It’s serverless, local, and browser based. You can load the latest binary from GitHub pages here:
Heh, me too. I'm on my third rewrite after a bunch of promising false starts.
Unfortunately geometric kernels are one of those things where unknown unknowns will bite you in the ass really hard because none of the content is really in the training data for LLMs and pathological/degenerate cases aren't just common but expected. IME it's not something that can be vibe coded with current models, if ever, without intimately understanding the algorithms.
I can't do a thorough review of waffle iron right now but just off the top of my head: it doesn't look like you have a tolerance context? The tolerances look like hard coded constants (TAU_MODEL/TAU_WORK/MATCH_TOLERANCE/etc) but that's fundamentally unworkable. Each operation and vertex/edge/etc needs to track accumulating errors and apply them to downstream point classification. Interfaces like Kernel::boolean_union(a, b) are the wrong abstraction because it's missing tons of information/functionality like accumulating FP errors, evidence/proofs, rollback, etc.
Thanks! I had a lot of false starts too. I asked fable in my repo about your question and it said there are two schools of thought on tolerance tracking. My kernel uses exact predicates which fable says eliminates the need to track tolerances the way ACIS does. I tried to paste the full reply but my comment seemed to have been auto flagged.
I think Fable gave you a load of nonsense. Like I said, this stuff isn’t really in the training data and academic research is disconnected from commercial CG (and I’m pretty sure its wrong about ACIS - iiuc ACIS used global tolerances and its one of the reasons it lost to PS).
Robust/(adaptive) exact predicates and tolerance tracking solve two different things and a geometric kernel needs both, especially if you plan on letting models use meshes, lattice, or freeform surfaces (which is where the industry is headed for simulation driven design). The constructions you feed the exact predicates need to already account for tolerances to be correct (i.e. pcurves, vertices, etc) and the second you do any math on them - before feeding it to the predicate - you need tracking. Predicates don’t eliminate the floating point math.
Six months. Basically when Claude code started to hit that inflection point I chose a project that seemed potentially impossible but was also something I care deeply about. (I have a lot of history with open source and CAD and I even made an open source CAD forum on GitHub a few years ago to discuss options.)
I set up remote tmux access to my phone right at the beginning so I have been advancing it rather continuously during my waking hours. Especially the last few months which have been a very focused push on a new kernel architecture.
The app consists of a wasm (web assembly) binary and JavaScript. The wasm runs locally on your machine in your browser and communicates with the JavaScript frontend. There is no backend server to handle any part of the program. The URL just loads the binary in to your local browser. I don’t know exactly how one would set this up but this would work for example fully offline. The browser basically just becomes a universal compute and rendering engine for it.
Author of the port here, you need a browser with JSPI support, which means recent Chrome, or Firefox Nightly with the feature flag flipped, or Firefox from the future.
I will also note that it's possible to compile this with wasm asyncify, but the result iirc is a ~400MiB Wasm binary that will crash the browser tab before you will be able to do anything useful in it.
I made the port, roughly ~one maxed out Claude Max 20x sub, at the bottom of the article I've shared the full claude code transcripts, so you can probably to some rough math on token usage with that.
Edit: to be precise 'maxed out' means one weekly limit on fable used over those 4 days
The launch graphic says "Ported end-to-end by Fable, an AI agent" and it definitely feels that way. Buggy, glitchy, needs some love and human eyes before it's really usable.
FreeCAD on desktop is much the same in my experience. If LLMs were mainstream a few years ago I would have assumed the UI was 100% AI generated with zero human input besides a oneshot prompt.
When was the last time that you used FreeCAD? Even 18 months ago, I would agree with you that it was a buggy mess, but I tried it a few weeks ago and it was super stable. The UI is still a little confusing, but that just comes with the domain, since AutoCAD and Fusion 360 aren't much better.
I couldn’t get the results described by the intro tutorial 3 months ago, in the parts bench drawing a polygon, snapping was broken.
If Solidworks and Onshape were born after the birth of LLMs, they'd probably be glitchy as hell.
Recent and related:
LibreCAD in the Browser - https://news.ycombinator.com/item?id=48755075 - July 2026 (17 comments)
Why would I want to run this in the browser vs locally?
Browser can be local. What’s nice about browser based is that browser based programs can run on every device. Though it sounds like this one requires chrome which seems weird to me.
Probably nice to have for those with low income that only have a Chromebook.
(I made this port) Fwiw I personally had no reason to do this port beyond using it as a benchmark of the agentic capability of Fable, where something of this shape is IMO a way better gauge than those dumb X.com 'I oneshot game with models X/Y/Z this is how it compares'
I published the actual prompts, and you can see quite clearly that vs Opus which is ok at implementing one big feature, Fable was really able to push through a good chunk of the port. That said it definitely didn't one-shot the port, it also didn't figure out a broken docker sbx sandbox by itself, and also later needed some gaslighting into thinking that the port is not really that hard (by any human measure it was quite hard given the scope of code involved.. The nearly 200MB wasm binary is mostly code afaict..). So there are some clear patterns of how the model was trained and also roughly the scope of task visible in those traces. What I see is that it likes prompts that would take an L4/L5 2-4 weeks to do with Cursor ~2 years ago, more needs some direction and deliberate prompting.
Because then it doesn't matter what you're running locally, as long as you've got a supported browser (Chrome, I'm guessing). It means it doesn't have to make a difference if you have a Window 10 desktop or a MacBook Air or a Chromebook. Go to the web page and look at this CAD.
This. The browser as a universal platform.
Like electron, but it's fine this way for some reason.
It is fine 'cause it uses the browser you're already using anyway instead of importing a new one for each and every application. Run the backend on a server somewhere and you can get to your application anywhere you can reach that server which - using a self-hosted VPN - can be anywhere you have internet access.
Onshape is free in the browser as long as you are not doing commercial work. It’s a professional system from the founders of Solidworks that competes with all the top CAD tools.
> It’s a professional system from the founders of Solidworks that competes with all the top CAD tools.
Because it's built on Parasolid, the same geometric kernel as everyone else. With ACIS pretty much out the door, almost all the professional CAD packages are just window dressing on the same CG implementation.
Agreed. I’m not OP but for six months I’ve been using Claude to build a from-scratch CAD kernel based on Rust and WASM, MIT licensed.
The actual UI still needs a lot of work, but I’ve been focused on the kernel. Fable has helped a lot though Opus was already making great headway.
I’m an OnShape power user going back about ten years, Solidworks before that. I need a CAD system that absolutely works. There is a lot of work to do still, and it still seems impossible to succeed, but I’ve been very happy with where things have been going with it lately.
It’s serverless, local, and browser based. You can load the latest binary from GitHub pages here:
https://sequoia-hope.github.io/waffle-iron/
Click the Assay menu to see the kernel test cases we’ve been using so far. Rapidly closing on 100% support!
Very cool. You mentioned it is mit licensed - is it shared publicly somewhere?
Heh, me too. I'm on my third rewrite after a bunch of promising false starts.
Unfortunately geometric kernels are one of those things where unknown unknowns will bite you in the ass really hard because none of the content is really in the training data for LLMs and pathological/degenerate cases aren't just common but expected. IME it's not something that can be vibe coded with current models, if ever, without intimately understanding the algorithms.
I can't do a thorough review of waffle iron right now but just off the top of my head: it doesn't look like you have a tolerance context? The tolerances look like hard coded constants (TAU_MODEL/TAU_WORK/MATCH_TOLERANCE/etc) but that's fundamentally unworkable. Each operation and vertex/edge/etc needs to track accumulating errors and apply them to downstream point classification. Interfaces like Kernel::boolean_union(a, b) are the wrong abstraction because it's missing tons of information/functionality like accumulating FP errors, evidence/proofs, rollback, etc.
Keep working at it! It's worth the challenge.
Thanks! I had a lot of false starts too. I asked fable in my repo about your question and it said there are two schools of thought on tolerance tracking. My kernel uses exact predicates which fable says eliminates the need to track tolerances the way ACIS does. I tried to paste the full reply but my comment seemed to have been auto flagged.
I think Fable gave you a load of nonsense. Like I said, this stuff isn’t really in the training data and academic research is disconnected from commercial CG (and I’m pretty sure its wrong about ACIS - iiuc ACIS used global tolerances and its one of the reasons it lost to PS).
Robust/(adaptive) exact predicates and tolerance tracking solve two different things and a geometric kernel needs both, especially if you plan on letting models use meshes, lattice, or freeform surfaces (which is where the industry is headed for simulation driven design). The constructions you feed the exact predicates need to already account for tolerances to be correct (i.e. pcurves, vertices, etc) and the second you do any math on them - before feeding it to the predicate - you need tracking. Predicates don’t eliminate the floating point math.
(Grain of salt and all that)
How long have you been building this?
Six months. Basically when Claude code started to hit that inflection point I chose a project that seemed potentially impossible but was also something I care deeply about. (I have a lot of history with open source and CAD and I even made an open source CAD forum on GitHub a few years ago to discuss options.)
I set up remote tmux access to my phone right at the beginning so I have been advancing it rather continuously during my waking hours. Especially the last few months which have been a very focused push on a new kernel architecture.
Respect.
Aww thank you so much! It has felt really good to work on as I have started to see something real come together.
> It’s serverless, local, and browser based.
uhh what..
The app consists of a wasm (web assembly) binary and JavaScript. The wasm runs locally on your machine in your browser and communicates with the JavaScript frontend. There is no backend server to handle any part of the program. The URL just loads the binary in to your local browser. I don’t know exactly how one would set this up but this would work for example fully offline. The browser basically just becomes a universal compute and rendering engine for it.
Sounds cool. Doesn't work.
Author of the port here, you need a browser with JSPI support, which means recent Chrome, or Firefox Nightly with the feature flag flipped, or Firefox from the future.
I will also note that it's possible to compile this with wasm asyncify, but the result iirc is a ~400MiB Wasm binary that will crash the browser tab before you will be able to do anything useful in it.
Seems it's only supporting Chrome at the moment.
Amazing. How much did it cost?
I made the port, roughly ~one maxed out Claude Max 20x sub, at the bottom of the article I've shared the full claude code transcripts, so you can probably to some rough math on token usage with that.
Edit: to be precise 'maxed out' means one weekly limit on fable used over those 4 days