About DuoCast

Hi, I'm Ivy Su (苏言)

I'm the founder of DuoCast. I spend a lot of my week commuting, and for years I carried around dense Chinese and Japanese PDFs — research papers, long newsletters, study material — that I never actually had time to read.

So I tried turning them into audio. Every text-to-speech tool I reached for had the same problem: in English it was passable, but in Chinese and Japanese it sounded flat and robotic — wrong tone, wrong pauses, no rhythm. Listening to it was worse than not listening at all.

The idea: two voices, not one

What finally worked wasn't a better narrator. It was two of them. When you turn a document into a back-and-forth conversation — one host asking the questions you'd ask, the other explaining — your brain stays with it. A monotone reading puts you to sleep; two people talking keeps you listening.

DuoCast does exactly that. Paste a document, a PDF, or a URL, and it rewrites the content into a warm two-host conversation, then voices it with self-hosted speech models tuned for natural Chinese and Japanese. Because we host the voices ourselves, there's no per-character meter — so a whole long report gets voiced end to end, not truncated.

What I care about

I care about how people actually retain spoken content, why a dialogue beats a lecture, and how good audio can make long-form reading something you finish instead of bookmark. I write about all of that on the blog.

If DuoCast helps you get through one document you'd otherwise have left unread, it's doing its job.

— Ivy Su, Founder & Audio Learning Researcher