Tesseract OCR 3.02 provides hin.traineddata for recognizing texts in devanagari scripts. However the Hindi training texts, images and box files are not provided, so it is difficult to improve the accuracy by further improving the traineddata. It is noted that recognition is more accurate and faster if the training is done with the same /similar font as used in the text to be OCRed.
I am experimenting with different fonts and training texts and will post the traineddata files for various devanagari fonts in the hope that these can be used to OCR the various scanned books with devanagari text.
Currently traineddata file for Sanskrit2003 font and another similar font used in a book are uploaded here.
See DocumentationWiki for more details.
