Interesting discussion, thanks for your inputs.

I didn’t find this pd4ml.page tag on the website, but the tag lib api did it (http://www.pd4ml.com/taglib/index.html). Looks like doing the job.

Regarding our stack design, it comes from one of our needs : to be able to generate mails with pdf attached to them, as a kind of newsletter. Indeed, we could have a request to process 1000+ of such mails, with just a few lines changing from one mail to the others.

in doing so, we would like to be able to spare the bandwidth as much as possible (and thus we plan to have a templating option) as well as be sure we won’t spoil some web server CPU (and as such being able to produce the pdf in a dedicated server). The templating stuff would be done before the pdf rendering engine.

hence our aimed architecture. I welcome any advice/feedback.

As PD4ML looks promising, I’ll now work on a demo implementation, using the trial version. Hopefully it’ll work out as expected and then we could discuss further of the buy and index stuff.

thanks again