Use the server-side recording (SSR) decoder service to generate .wav files from the audio data that is captured by SSR and then uploaded to the cloud as .tar files.

After the .wav files are decoded, they are compressed into .zip files and uploaded back to the cloud in a subfolder of where the .tar files were found. By default, decoded subscriptions result in a single .wav file for each participant in the recorded conversation, and all files are the same duration. Other options are detailed in the SSR API documentation.