Nekudot

Description

If you're like me and learning Hebrew, but don't yet know how to pronounce new words without nekudot being present, this tool is for you.
With this extension, if you highlight the Hebrew text and click on the extension's icon, Nekudot will process that text through a machine learning model and replace the Hebrew right on the web-page to include nekudot.
This isn't guaranteed to be perfect, but it performs really well especially if you are looking to pronounce Sephardi, or Israeli, Hebrew where Kamatz and Pataḥ sound the same.

In addition, this is all free and open-source at github.com/GiladAmar/Nekudot. So if you're looking to learn how to build a Chrome extension feel free to take a peak. In particular, using TensorFlow.js to work with machine learning models in the browser.

Nekudot is hot off the press, and so I expect there to be issues I haven't seen. If you run into any, such as text becoming jangled, or elements disappearing, please let me know how to recreate the issue. For example what website and exactly from where to where you had highlighted. A screenshot couldn't hurt. That will help me diagnose and fix the issue. A good place to do this is on the repository issues tab here: github.com/GiladAmar/Nekudot/issues

Known issues:
* "I'm not seeing a Nekudot icon where the extensions are"
This is likely because chrome has elected not to show it. You can change this by clicking on the puzzle-piece shaped icon on the top right and pinning the Nekudot icon to the toolbar.
* "Israelis laugh at me when I try speak Hebrew"
Sorry, can't help you there.

Upcoming features:
If you have any suggestions for features don't hesitate to contact me via the Store or on GitHub.

If you're in a generous mood, or simply have too much money, and have gotten value out of this tool I'll happily accept a donation and take it off your hands at paypal.me/GiladAmar.
If not, no worries, it's free and will always be.

Last but not least; this wouldn't be possible without the work done by Elazar Gershuni and Yuval Pinter in arxiv.org/abs/2105.05209 who graciously let me use their LSTM model. Thank you, it's a really cool and original use of machine learning.