India's 22 scheduled languages are used by more than a billion people in conversations that frequently mix languages, sometimes within a single turn, and often overlap freely. These exchanges take place over telephones, in meetings, and in media, making it difficult to determine who spoke when.

The paper introduces Anuvani-Diarization, a zero-shot method for speaker diarization across these languages, along with proposed annotation conventions. The zero-shot design aims to work without language-specific training data, which is important for languages with limited speech resources.

The abstract does not include experimental results, so the effectiveness of the approach is not detailed in the source. The contribution appears to be both a method and a set of labeling guidelines to support future work in multilingual speaker diarization.