Flume replicate then filter, how?

I need to replicate a source, then filter only one of the replicas on a regex.

For example, if my source data was 'dog, donkey, cat' I'd like to sink 'dog, donkey, cat' to one place, but sink only 'dog, donkey' to another place (d*).

I'm planning to use the regex interceptor but I'm not seeing how to do this since the interceptor is associated with the source. How could this be accomplished?

Had to make another flume agent tier for this. Replicated on current agents, and forwarded to new flume agent processes where it could attach the regex interceptor to the source.