Despite jamie already providing impressive summaries in some scenarios, we often got the feedback from our users that the quality of summaries could be improved. Hence, we started to work on ways to improve the quality of summaries. One of the core elements we believe will have a tremendous impact on the quality of summaries is identifying who is saying what. With this release, we are excited to ship the first implementation of speaker diarization. This allows us to detect different speakers in a meeting based on their voice. This way, the transcripts we use for the generation of summaries are much more structured.
We all know that most of us are using headphones while being in calls. That's why we are excited to ship headphone mode to jamie with this release. This allows jamie to capture the system audio output together with the microphone. This way, you can finally start using jamie even if you are using headphones. We start off by shipping this to Windows and will add it to macOS soon.