Runs buzzdetect's models on audio, the same files the desktop app uses. All processing happens locally on your browser; audio isn't uploaded.
Pick a model beside the Neurons button. fast_v1 is quickest but most prone to false positives; heavy_v1 is slowest, the most resistant to false positives, and offers the most neurons (it's a 21 MB download); balanced_v1 sits between them.
fast_v1
heavy_v1
balanced_v1
MM:SS.SS
Pick which neurons you want to plot and set a detection threshold for each.