Maybe too late, but I stumbled over this question 2 weeks ago and found a solution by doing some coding:
I managed to solve a similar issue with bulkWaterMark and .NET, where you can watermark dynamical text via a feature called expressions. I used the Freeware version of the tool and coded a custom expression context plug-in that loads data from an xml file where image informations are stored. Then I wrote an expression that takes the current image's filename as parameter and looks it up in the xml file - all done in C#.
The problem was that there is no SDK documentation available and I'm also not sure if this plug-in interface is officially public, but you just have to inherit 2 or 3 classes and override some methods to inject your own code into the application. It worked like a charm then.
You just need to reference the assembly PMlabs.GrfX.Framework and create your own implementation of these three classes:
The Expression class is the function that is called when an image gets rendered. The context holds multiple expressions and the plug-in class holds the context. Just make sure to implement everything in the abstract Expression class, the others are just plumbing code.
After compiling your plug-in assembly put it into the bin/plugins folder of your bulkWaterMark installation.
Another edit: Published a simplified code of my solution at Github today: https://github.com/barnacleboy/bulkWaterMarkExpressionPlugin