If someone is blind, then they are likely already using text to speech software like JAWS or NVDA.

Considering the sheer size of wikipedia, I can't see much point TBH.

Screen Readers already do this. I have the Wikipedia app on my iPhone and use VoiceOver which has adjustable and very natural voice feedback in several languages. Wikipedia in a desktop browser is perfectly usable as well with VoiceOver on my Mac, again using the same synthesized high-quality voices sped up to the rate that I want to use and can audibly parse. So unfortunately your idea wouldn't be used nor inherently useful to blind users since we are already using our own technology to do exactly what you are proposing.

I'd personally not use it as I can already use wikipedia, sure yearing left parron link 1 right parron, or however it's said can get a bit annoying but you get used to it.

Very thoughtful of you but not honestly very practical of an idea and there’s already multiple ways to do this. I honestly think you’d be wasting your time. Is this for a class or something nice and generous you would like to do. I suggest you invest in something more needed. I would ask around for need if it’s a project.

Wikipedia does have a spoken article database, though obviously it doesn't include all articles. https://en.m.wikipedia.org/wiki/Wikipedia:Spoken_articles

Actually, to a degree this already exists. Anyone with Amazon Alexa can ask her to Wikipedia whatever and she reads it. It is a bit clunky, asking every few paragraphs if you want to continue and Alexa’s voice may not be your thing, but it is a start....