We have been using an OCR utility in Confluence server version 6.8 to render non-OCR files searchable. The output from the OCR utility is a file with the same name but with .ocr.txt. For example, in confluence version 6.8 a non-OCR file like filename.pdf will be uploaded. The OCR utility will OCR the document and save filename.pdf and a companion text file named filename.pdf.ocr.txt