Hi Team, I'm getting "message.duration" as 0 in the transcript response when trying in Korean/Japanese language with "Cartesia" for Voice and Transcriber.

Based on the your report that the "message.duration" is returning 0 in the transcript response for Korean/Japanese languages when using the Cartesia Voice and Transcriber service, the most likely solution involves checking for a service-side bug or integration issue specific to how Cartesia processes duration metadata for these CJK (Chinese, Japanese, Korean) languages. The immediate actions should be to verify the API's language codes are correct (e.g., ko, ja), confirm the input audio format and quality aren't corrupting metadata, and, most importantly, report the issue to the Cartesia engineering team as this points to a fault in their internal audio segmentation or duration calculation pipeline for those specific language models, which is often a known issue fixed in service updates.

Hi, could you please provide a call id where you experienced this issue? We would like to review the logs to identify the cause.