This is terrific! I listened to it for about a half hour last night, and it was far more useful than I thought! One of the articles I had read earlier in the day, and your parser did a great job of using different voices for different people quoted in the article, which made the article flow in a way that's pretty close to real speech.<p>One of the articles, however, seemed to switch from voice to voice for no discerning reason. It might have been a medium article, where authors are more likely to stylize the text with bold, underline, italics, etc. That made it a bit confusing and difficult to follow.<p>For me, listening to any spoken word requires a little mental effort. For example, listening to an NPR podcast takes about 10%. Your app maybe took 20%. That's within a margin that, if you're able to smooth out your speech algorithms and tune the voice-choice algorithm, I could imagine this being as easy to listen to as a podcast.<p>So, on the whole, the voice-choice algorithm needs a little work, but other than that, I really like this!