ARM64 support

[mkattut]

(I don't have much experience with Github so, if I'm reporting this incorrectly, please let me know the correct way)

I recently received a laptop that runs on a Snapdragon X Elite chip instead of an Intel chip, and I recently attempted to use MathCAT 0.7.5 with NVDA 2025.3.2 to read math content for a presentation. The math goes unread in Chrome and Edge (regardless of math renderer used) but is read in Mozilla FireFox.

I asked Google Disability Support for any insights they had. They requested additional information, which is included in this video of my experience.

They then came to the following conclusion:

The issue appears to be related to the architecture of your new laptop (ARM64, Snapdragon X Elite). While the latest version of NVDA (2025.3.2) supports ARM64 devices, the MathCAT add-on currently contains components designed only for Intel/AMD processors (x64/x86). Because of this mismatch, NVDA cannot load MathCAT when using Chrome or Edge, which is why the math content is not accessible in these browsers.

Firefox handles math content differently and does not rely on the older MathCAT components, which is why it works as a reliable workaround.

Our internal team recommends reaching out to the MathCAT developers to request ARM64-compatible components. Once this support is available, it is expected to resolve the issue in Chrome and Edge.

We understand this may be inconvenient, and we appreciate your patience as this technical limitation is addressed. Please let us know if you need guidance on contacting the MathCAT team or if there is any other assistance we can provide.
[...]
Thanks!
Jaypee
The Google Support Team

[7-4091000040372] Your Google Disability Support Inquiry

I may be limited in my ability to provide additional information, but let me know if you need more information.

[NSoiffer]

Thanks for your report. Indeed that will be a problem. Thankfully it should be relatively easy to fix at least on personal level (I point you to a file to install). I'll need to reach out to the NVDA team to see how they solve this more generally. I suspect that there are a few other components that are compiled for specific architectures.

I'm currently facing a deadline of Monday to get a couple of papers turned into a conference. I'll look into the ARM build next week. If you don't hear back from me in two weeks, please re-raise this so it comes up to the top of my queue.