nCompass provides an audio stream denoiser that focuses on removing voices in the background. With this, it is now possible to transcribe audio with multiple overlapping voices. We also provide this in realtime for cases that require live audio denoising.

In a world where conversations are increasingly virtual and our most powerful machine learning models are optimized for text inputs, having comprehensible transcriptions even in the precense of background voices is vital for developing more intelligent audio solutions.