A friend of mine works in e-commerce and often needs to edit product videos that already have text or subtitles burned into them.
I tried a few existing tools and APIs and found that Alibaba's VideoDetext worked reasonably well for the videos I tested, so I built a simple web interface around it.
The workflow is pretty simple now: upload a video, detect or select the text area, and remove it.
It's still an early version. The biggest issue I'm seeing is that complex backgrounds can sometimes become slightly blurry after the text is removed, so that's what I'm working on next.
I'd be interested to hear how other people are handling this kind of video editing workflow.