We just shipped a backend update that should fix issues some users were seeing with disfluencies (ums, ahs) in the transcript, as well as names not being corrected properly. Please let me know if you’re still seeing these problems and send me examples. Thanks!