pipe workflows or data through rapidminer

I'm looking for easy ways to automatically process incoming data as easy and fast as possible.

So I like to know if there is a way to use rapid miner in a pipeline mode with another
program which pipes new data to rm.

Another possibility would be to let rm set a trigger on a data base or observing a
file or a directory and automatically read and process incoming data.

The aim is clearly to avoid the startup of the rm environment which takes rather long.

Thanks in advance,

Michael