Recently, I was playing around with speech AI tech trying to dub some anime for myself (lol), and I found the quality of dubbed videos from existing products wasn’t good. So I try to build one myself.

After a few weeks of hacking, it’s finally at a good enough stage to be released. The whole pipeline is not easy to build, with lots of moving parts with different models. I plan to continue optimizing it.

Give it a try and let me know your feedback. Thanks.

Demo: https://www.youtube.com/watch?v=JIlz4VUmeBk Product: https://trydub.com