exam-listening-verbatim

Upload the audio file and request a verbatim transcript. The transcription preserves exact wording, repetitions, filler words, numbers, and spelled letters, and any segment that cannot be verified after re-listening is marked in place rather than guessed.

Provide the exam audio and ask for a full verbatim transcript. High-risk regions such as fast speech, accents, proper nouns, dates, and prices are re-listened to carefully, and file boundaries are checked so no opening or closing lines are missed.

Yes, timestamps and speaker separation can be added when requested, while the spoken content itself stays verbatim. By default, speaker changes are shown with line breaks and labels like Speaker 1 are only used on request.

Unverifiable segments are marked at their exact position with [không nghe rõ] or [không chắc] instead of being filled in from context or exam templates. You can then send a clearer version or a shorter clip of that segment for re-transcription.

No, the default output is only the verbatim transcript in the original audio language. Translation, summarization, or answers are produced separately only when explicitly requested, never mixed into the transcript.