Voice Profiles
Voice profiles enable speaker identification in audio and video content. The system learns to recognize who is speaking, automatically attributing spoken content to the right person.
How it works
When audio or video is transcribed, speaker diarization separates different voices. Voice profiles can be linked to people profiles, so the system recognizes the same person across different recordings. Over time, voice identification improves as more samples are associated with each profile.
Transcription with speaker labels linked to people profiles
Why it matters
Knowing who said what transforms transcriptions from anonymous text into attributed conversations. In meetings, interviews, and recordings with multiple speakers, voice profiles ensure every statement is linked to the right person. This makes it possible to search for 'what did Alex say about the timeline' and get precise results.
Search result showing attributed speech from a specific person