Ocr api questions

Hello I am wanting to find out if the ocr.space api can detect text of locally stored video files without the need to have the video playing. If so, can it detect digital text as in subtitles or from an overlay and also text from in video content for example text from a sign that a person in video is holding?

i am also working at image text converter API. It is also working littlle but glitchy can someone help?