Skip to content
Chad edited this page Sep 21, 2017 · 5 revisions

Welcome to the minutes wiki! 🎀

minutes is a standalone API for audio speaker diaratization.

Speaker diarisation (or diarization) is a process for partitioning an input audio stream into homogeneous segments according to the speaker identity - Wikipedia.

πŸ‘‰ Input

As input, minutes accepts an audio recording of a conversation including n speakers, and n labeled audio samples (one for each speaker).

πŸ‘ˆ Output

As output, minutes produces a list of phrases from the conversation - each phrase should have the following keys: speaker, start_time, end_time and body.

Clone this wiki locally