Running the dashboard aggregator at 03:00 PHT and the transcript sync job keeps choking on extended calls. Trying to pull the complete voice-to-text output via GET /api/v2/analytics/speech/text with the conversationId filter. The first request returns fine but the segments array cuts off after exactly 42 entries. Checked the Link header for pagination logic but the next cursor points to a dead route. Switched to the Python SDK method get_speech_text_analytics with interval_grouping set to 1d and it’s still doing the same thing. The response structure looks like this:
{
"segments": [
{"start": "2023-10-25T08:00:00Z", "end": "2023-10-25T08:05:00Z", "text": "..."},
{"start": "2023-10-25T08:05:00Z", "end": "2023-10-25T08:10:00Z", "text": "..."}
],
"count": 42,
"next": "/api/v2/analytics/speech/text?cursor=xyz789"
}
Hitting that next endpoint throws a 404 Not Found. The docs mention a limit parameter but it caps at 1000 and still drops the tail. We can’t chain these requests without the SDK overwriting the bearer token. Anyone figured out how to handle the cursor pagination for long transcripts? The aggregator just hangs on the final chunk.