This is tricky to do with free software, but it can be done. PDFBeads
can make tiny PDFs with OCR text layers.
One of the problems with making PDFs is that compressed images in greyscale or colour are huge - very huge. It's very difficult to get them down to an acceptable size. That's why many people use Scan Tailor. It reduces pages to pure black and white, which compresses very well. With the right tools, you can make books of hundreds of pages into a PDF under 10MB, with OCR.
The other problem is that not all OCR tools give you an output that's suitable to be used by other software this way, so PDFBeads requires that you use the Tesseract or Cuneiform open-source OCR programs.
The opinions expressed in this post are my own and do not necessarily represent those of the Canadian Museum for Human Rights.