02 / REVIEW
Transcript
Listen & correct, one segment at a time
AI Proofread sends only transcript text and short vocabulary references to OpenAI. Audio goes only to your configured ASR service; OpenAI receives text only. Review and accept suggestions before applying them. Export aligned JSON to save originals and edits. ASR segments are starting points, not guaranteed sentence boundaries. Adjust times; split at the audio playhead and text cursor, or merge with the next segment.
Uses the existing deterministic segmenter.
Ready. No audio is processed automatically.
03 / ANALYSIS
Communication breakdown
Not analyzedAutomatic extraction is a review aid. Conformance checks MTP syntax, not ASR accuracy or equivalence to the recording.
Your extracted command and communication stages will appear here.