The Clean view is editable — tweak anything before downloading. Markup compares the current text against the original.
IntervalInsert timecodes inline at a regular elapsed-audio interval.
Every
Speaker labelsTimestamp only when the speaker changes, directly before the existing label: [(16:23)] Interviewer:
Timecode offsetUse normal HH:MM:SS, or frame-aware HH:MM:SS:FF. For frame offsets, choose the source FPS manually.
:::
FPSChoose FPS manually for frame offsets
All output timecodes use total elapsed minutes and seconds: [(01:16)], [(67:33)], [(103:16)]. Speaker labels are read from the TXT itself and are never generated by AssemblyAI.
Preparing…
Timecoded transcriptReady to review
Alignment
Precision verification
Waiting for acoustic verification.
Precision
0verified
0auto-adjusted
0needs review
—median model difference
Questionable timestamps
transcript_timecoded.txt
The output preserves your uploaded TXT and only inserts company-format timecodes. Speaker-change stamps sit directly before the existing speaker label.
Preparing private upload…
Verbatim transcriptRev AI transcript ready
Rev AI
transcript_verbatim.txt
Rev wording is preserved. Original media goes through the same private R2 + Rev AI asynchronous workflow. Rev timestamps are removed, speaker turns remain labeled and separated by blank lines, single digits 0–9 are spelled out, and sentence-starting “So” receives a comma. Um/uh and other disfluencies are retained.