Honestly I use this to help me speed read long books. I will normally have a physical copy in hand and pull up the web version and have Read Aloud running at 1.5x I have to read a lot by virtue of my interests and this helps, but it is certainly more useful for articles and uninterrupted prose like fiction.

I read tech/programming books. I wasn't expecting this to do a good job reading all the symbols and concatenated words but it does a relatively decent job. Some symbols are ignored or not read correctly but that's to be expected.

What's annoying is that the parser for this extension frequently skips single sentences or footnotes that are in between two larger paragraphs. So like if you have one paragraph then a sentence from the next paragraph followed by a diagram then the rest of the paragraph Read Aloud will skip straight from the first paragraph to the larger chunk of the second paragraph. sometimes the skipped part is important so i find myself pausing and rewinding frequently to make sure i got everything. It's jarring when you have a book in hand and read and hear two different things.

Feature suggestion: Consider highlighting the text that is already on the page by default. And having the option to use the modal if the page has funky formatting or something. Is there any reason why you can't just manipulate the DOM to highlight & scroll the words that are already on the page?