Looking for advice on a discrepancy in recording metadata when exporting WebRTC softphone interactions for a legal discovery request. We are using the Genesys Cloud v2 API with a Python 3.9 client (purecloud-platform-client 142.0.0) to trigger bulk export jobs to an S3 bucket. The issue arises specifically with digital channels routed through the WebRTC softphone. While standard voice calls export correctly with full chain of custody metadata, the softphone recordings often lack the interactionType field in the resulting JSON manifest, or it defaults to “voice” instead of “webRtc”. This breaks our automated legal hold indexing process, which relies on strict metadata schemas to map recordings to specific agent workspaces.
The environment is set to Europe/London, and we have confirmed that the Architect flow correctly tags the interaction as a WebRTC session. However, when the bulk export job completes, the metadata payload for these specific files is incomplete. We are observing a 200 OK response for the job status, but the downstream validation fails because the recordingType field is missing or incorrect. We have checked the rate limits and permission scopes, and they are identical to the working voice call exports. Is there a known limitation with WebRTC softphone metadata propagation in the bulk export service? Or is there a specific filter parameter we need to include in the POST /api/v2/recordings/jobs request to ensure the softphone-specific attributes are captured in the S3 manifest? We need to maintain an unbroken chain of custody, so any workaround or API update information would be critical for our current discovery timeline.