Hey folks,
Trying to build a quick tool to toggle call recordings on demand. We’ve got some sensitive calls where we need to pause recording for legal reasons, then resume later.
I’m hitting the POST /api/v2/recordings/jobs endpoint with the following JSON payload:
{
"name": "Manual Recording Job",
"type": "conversation",
"mediaType": "audio",
"conversationId": "12345-abc-def",
"recordingType": "manual",
"state": "started"
}
The API returns a 201 Created with a job ID. I can see the job in the recordings console as “In Progress”. But here’s the kicker - the agents’ audio isn’t actually being recorded. The transcript shows gaps, and the final recording file is silent for those segments.
I’ve checked the permissions on the OAuth token, and it has recording:write. The conversation is definitely active. Anyone else hit this wall with the Recording API? Am I missing a header or a specific parameter to actually capture the audio stream?
Checked the docs, but they’re pretty light on the manual recording use case. Feeling stuck.