this is a tool I built for fun two years ago, which I now decided to release as open source.
It's a Java program that takes a Wikipedia page as input creates a video narrating the content of the page using a combination of sources for images and videos, plus Amazon Polly for speech synthesis.
At the time, I wrote a detailed explanation[1] of how the tool works, if you're interested.
The project is not active or currently maintained - I'm just releasing it because a few people requested it to me, and so I thought I might as well post it here!
EDIT: Forgot to specify: although, in theory, this works with any Wikipedia page, it might not do so very well (or at all) in a few cases. For example if the input page is very short or it has an ambiguous word which will then not result in good matching images from Pixabay.