Transcription Modes

VoxParse supports two transcription modes. Both cost the same and return the same structured AI analysis — only the transcript_cleaned field differs.

Mode	Behavior	Best for
`polished`	Removes filler words (um, uh), false starts, stammering. Fixes ASR errors and produces clean, professional text.	Business reports, CRM notes, dashboards, customer-facing summaries
`verbatim`	Preserves all filler words, hesitations, self-corrections, and false starts. Only fixes obvious ASR misrecognitions.	QA review, compliance audits, dispute resolution, agent coaching, training data

Same price: Both modes are available on any plan (Pro or Lite). See a side-by-side comparison with real audio