I have just incorporated PDFTechLib into one of our projects. We are using it primarily to modify existing PDFs which come in from customers, including adding new text, images, lines, boxes and barcodes. In order to decide how to process each PDF, we're extracting the text from the original PDF to see what it contains. We are also using the library to create simple PDFs from scratch, to merge pages from one PDF into another, and to delete pages from existing PDFs. There are facilities for automatically flowing text onto the page which look to be very useful for creating more complex PDFs, but we don't need to use that facility at the moment.
The library is quite straightforward to use. I haven't tested it yet with very large PDFs but the performance seems fine so far. A little more detail in the documentation would be useful, but so far I haven't come across anything where I couldn't actually work out how to use it.
Support has been first class. I found one minor problem which was fixed within a couple of days and a new version released for me to test. The developers are also very receptive to suggestions for enhancements and several of mine were incorporated within a few days. For example, I asked to be able to extract text from a specified area on the page in addition to being able to extract the entire text, and this has been added which makes it easy to extract page numbers, invoice numbers, and so on. A couple of my other less important suggestions have also been added to the list for future development. This whole process has given me confidence that if I should come across a show-stopper problem at any time, there's a good chance a fix will be produced rapidly unlike some other suppliers I have dealt with, so I'm now happy to use the library in production software.