Use Voice Recognition / Machine Learning from YouTube Live Stream to Identify Page Number (Voice to Text Recognition) + other features

Part 1: Churches to display their bible side by side to a live stream of their church. When the church leader or anyone from the live stream says "Turn to page 52" or "page 21" then the PDF viewer would automatically turn to that page. This would need to utilize voice recognition and/or machine learning to better improve over time. Part 2: There would be a slider function for each viewer to be able to adapt to the size they need whether they want to focus on just the video or really zoom in on the bible. Part 3: There needs to be a simple backend on the website (or software) where the user inputs a YouTube Live Channel Link + Google Drive PDF LInk + [login to view URL] (Live Chat Interface). All of this would create a single embed code where each church can place on their website. This would all for the live stream, PDF, and live chat to all be placed on their site with all the functionality listed above. HERE IS A ROUGH EXAMPLE I CREATED SHOWING THE DESIGN: [login to view URL]