Create a script that will click on the extension button in the browser.

Hello. My programmer thought about automating this through the backend using Selenium. But he looked at it and said it would be better, faster, and easier to use a script in Js. I have a program that is built on the backend and frontend. The page runs locally in the browser. On the pages with the Excel file, for convenience, an article is loaded, and this is done using HTML, CSS + Js script. And sometimes in the browser itself, various extensions are used, like grammar checking and so on. When there is one article, it is not a problem to click the button once. But when I load 10 - 50 such articles, I think it's clear. 
Task for the script. 
We add a button to the page. Where the articles are loaded. There are many other buttons of mine from which we can take style copies and so on. And clone one of the buttons for the future button. Now the script itself. What it needs to do.
1. Click the extension button. The button itself appears floating in every visual block where there is an article. The script needs to click on it. One by one. And here is the most important point. It must understand that until the extension has completed its action, it should not go and click on another button in another block. We thought about tracking the response in the console, in the network, but there is nothing written there. There is no response code. So here we will have to latch onto something else. Or wait for a timing, for example, 10 seconds, which is very trivial. Or a more precise option. The script observes how many characters are in the block. For example, there were 1000. After the action of the extension, for example, it checked and corrected errors. The number of characters will change in the block. For the script, this is a signal that the extension has completed its work in this block. And it can go to another block and click the button there. 
2. There can be a lot of text in the block. And the extension may write a standard error. With text. How I manually deal with it. I highlight half of the article's text and click the button again. The script needs to have the same logic. If there is an error, highlight half of the article and click the button again. There are different ways to approach this. And how can the script know how much half of the article is? It's simple. Either a percentage of 100 is 50. Or the number of characters from 1000 is 500. That's it.