Is it possible? With no library's or memory of the changes as they're made?

The example I'm showing below is probably more edited than the text I will be evaluating (although much shorter). For my use case, I will be editing a transcription that is usually very accurate and I need to be able to know if a new word is added & where so I can approximate a timecode for the new word, as well as shift the existing timecodes forwards to combat the offset that the new word has created.

originalString = "Hello, this sum example txt. Hopefully this is possible!";
editedString = "Hello, this is some example text. Maybe this is not impossible.";

//The edited words are: "sum/some", "txt/text", "Hopefully/Maybe", "possible/impossible"
//It would be useful to get the edited words also - but not essential

//the new words are: "is", "not"
//Output would be
newWordIdxs = [2, 9];

First post here and have only been coding for 3 or 4 months so any tips on how to ask better questions are very welcome.