online atsa

[boxofbox]

Is there a way to create a stream/buffer like analysis (with minimal delay) so that ats can be run on live audio?
would benefit from optimizing the dropping of tracks (limiting search space) and improving default sorted state.