ht://dig works very well, although it works best with an external parser included with the XPDF package. There is a nice perl script included with it which extracts the document information from the header portion of PDF files as well. If you are using ht://dig, make sure you set your maximum doc size to the size of your largest PDF file. This may take longer, but it will otherwise abort on that particular file and just move on (it can't parse partial PDF's).
ht://dig can also be directed to export its word database to an external, parseable file for other program's use...you could certainly use PHP to then wade through that file, but you wouldn't get quite the other features of ht://dig in the process.
Check out the documentation on www.htdig.org.
