Hey everyone,
I’ve been diving into the GC audio APIs lately, exploring how we can leverage recording analysis for voice biometrics workflows. I’m hitting a snag when trying to retrieve detailed recording data. When I filter the media type to audio in the Tokyo production environment, the request returns a 422 Unprocessable Entity.
I also noticed that after the v2.8.4 SDK update, the console drops the audio duration metric entirely.
To keep things moving, I’ve adjusted the request to bypass the filter and am now pulling raw PCM streams via the export jobs. While this workaround secures the audio stream access, the processing lag is substantial. It’s really hurting the downstream fingerprinting pipeline and blocks any future speaker verification rollouts I have in mind.
Has anyone else run into this with the audio filters in Tokyo?