Stuck on a discrepancy between the live WebRTC session logs and the metadata provided in the subsequent bulk export job for legal discovery purposes. We are operating in the eu-west-2 region using Genesys Cloud v12.9 with BYOC S3 storage enabled. When agents initiate a call via the native WebRTC softphone, the interaction is correctly tagged as digital in the UI, but the recording object returned by the GET /api/v2/recordings/jobs/{jobId} endpoint lists the mediaType as voice rather than web_rtc. This classification error propagates to the CSV manifest file generated during the export, causing our chain-of-custody validation scripts to flag the files as missing digital channel metadata. The specific error observed in the application logs is a 400 Bad Request when attempting to filter exports by mediaType: web_rtc, with the message Filter criteria not supported for selected media type. We have verified that the recording API scope is granted to the service account used for the export job. The issue persists across multiple agents and does not appear to be related to browser version or codec negotiation. The interactionId links correctly to the conversation history, which shows the channel as web_rtc, yet the recording job itself seems to inherit the voice configuration from the underlying telephony trunk. Is there a specific parameter required in the POST /api/v2/recordings/jobs request body to force the correct media type classification for WebRTC interactions? We need to ensure the exported files retain the legalHold flag associated with digital channels for compliance. Any insight into whether this is a known limitation of the current bulk export logic or a configuration issue within the telephony provider setup would be appreciated.