Thanks for the suggestion.
As we wrote in email, the problem here is that every free-form text file written by people is written in a different way. Some people start with the title, and then the instructions, and then the ingredients. Some people interweave the ingredients in with the instructions. Some people mis-spell important fields that start or end something. Some people write the "Yields 6 servings", and other people say "Serves 6-12". etc. For this reason, no recipe software has yet figured out a way to fully interpret written English and parse through free-form text files to extract recipe content; in the meantime, the best we can do is the point-and-click Screen Import process.
However, if ALL of your text files are written in a very consistent manner, with the same delimiter (i.e., the same end character at the end of each field), perhaps in the long run there's a way we can let you "teach" the software what the precise format is of each recipe.