It works with texts, so it support lots of languages, but now I'm using the sentence tokenizer only for english, so in other languages that sentences are not correctly detected. It should be improved in the future, but I realized, I don't need 100 % correct detection of sentences. I'm using OpenNLP libraries for it. It scrapes webpages with Jsoup. Again I'm using libraries like Readability4J and Crux for parsing the webpages and removing ads etc.. So, the main idea and payoff of the project is to have a chunk of text associated with the picture. It's good for focus and have distraction free environment. For example, for me, it is a problem to read longer texts and don't be distracted by the surrounding chunks of text (before and after the chunk I'm reading in a moment). I have a positive feedback of one dyslexia patient, but he is no longer on Facebook, so it's the question if it helps dyslexia as well. But I personally use it for incremental reading webpages almost daily and I can better focus with it. I'm glad to share it as open source.
I was working on this project https://github.com/lucidl/tededroid before, now I'm going to record a video with usage, I'm looking for suitable article with text and lots of pictures.