This looks like a great tool. Whilst YouTube viewers usually have their sound on for videos, on other social networks, studies have shown that a large percentage have their audio muted whilst watching/scrolling through videos on other networks like Facebook, Twitter and Instagram. So being able to add these captions will add a whole new dynamic.
The only area I can see this being an issue is where the video has a background music/song, as this would cause the system to struggle to pick up the voice correctly and provide accurate captions.
Overall, a very exciting product and I look forward to trying it out.