When converting YouTube articles to text, it might be incomplete, only converting part of it while the rest is not displayed. This area needs improvement, keep it up! In videos within the same file, AI sometimes doesn't know the content of the video, which can lead to irrelevant answers. Could there be an option to allow multiple choice selections of resources, enabling users to ask questions related to the selected material?