GET /api/v2/conversations vs /api/v2/analytics/conversations for bulk extraction

Is it possible to reliably fetch raw interaction records for Snowflake staging?

Context:
I executed GET /api/v2/conversations. Response: 403 Forbidden. I pivoted to GET /api/v2/analytics/conversations. The output aggregates metrics rather than emitting discrete JSON objects.

Question:
Which route handles bulk extraction of unaggregated conversation data? My Python pipeline requires exact start/end timestamps and direction flags, not precalculated SL values.