I would like to create an app which can scan documents, find all text in the document and analyze the document. The documents are invitation of tenders for windows in construction. The windows have standardized parameters which are spread through the document. The app should have the ability to be trained to connect the parameters and give a full set of parameters as an output f.e. parameters would be "width", "length", "type of glass", "type of frame", "sound insulation", "colour" etc. The parameters are listed in different places of the document. The user should then train the app to connect the parameters to a set of parameters for one window.