I have documents which use only one font throughout the document. Different documents might have different fonts, but I know which document uses which font.
No, I don't think Tesseract supports such an option. What you can do is to train for one specific font and then specify that traineddata during recognition of your documents.