V‑Spark processing begins with audio data, but the eventual database record entails much more. For V‑Spark, the phrase audio record denotes that database record, which includes the initial audio data along with its JSON transcript, metadata, and analytics data. This record may be referenced by two attributes:

• requestID — generated when a file or batch of files is first uploaded.

• transcriptID — generated when a single audio file is submitted to the ASR engine for transcription.

Records for audio files submitted together share a requestID, but individual audio records always have unique transcriptIDs.