I wonder whether the Tika Language Detector works completely off-line, thus being able to guarantee the privacy of the processed text, or whether there is some kind of on-line matching or telemetry data exchange happening in the background ?
From what I could find out here, I assume that the process could be done completely off-line.