Hi Zoom Developer Community,
I’m currently investigating inconsistencies in transcription availability for Zoom Phone calls and would appreciate clarification on expected behavior.
Questions
-
Is transcription strictly dependent on cloud recording availability?
Or can it be generated independently like AI call summaries? -
Are there minimum duration or audio quality thresholds required for transcription?
-
In multi-leg call flows (queue → extension), where should transcription be expected?
-
At the queue recording level only?
-
Or should it propagate to the final call leg?
-
-
For PSTN calls, is recording (and therefore transcription) expected to be disabled by default?
-
What are the exact conditions where:
-
recording_idexists -
ai_call_summary_idexists -
BUT transcription is still not generated?
-