Advanced Proof listens to your entire audiobook against the manuscript and flags every misread, skipped and added word — with timestamps. You review each catch, and the pickup files for your narrator build themselves.
Upload your manuscript and your chapter audio — MP3, WAV, AIFF, M4A or FLAC. Advanced Proof does the listening; these are the four things it hands back.
Misread words, skipped lines, added words — each finding shows the manuscript line against what was actually narrated, with the exact timestamp. Play the moment, then accept it for re-record or dismiss it. The final call is always yours.
Every accepted re-record lands in a numbered pickup script — chapter, timestamp, the reason, and the full manuscript sentence for your narrator to read. In book order, ready for the session.
Alongside the script comes a preview audio file: a beep, then the original narration of each pickup, in the same numbered order — so your narrator matches the tone, pace and room of the original take.
A downloadable report of every mark — timestamp and chapter in separate columns, ordered by book position, with the action, the quoted line, the reason and where it came from. Built to sit next to your editing software.
| Time | Ch | Action | Line | Source |
|---|---|---|---|---|
| 07:41 | 3 | Re-record | “the carriage waited” | AUTO |
| 22:15 | 7 | Re-record | “very quietly” | AUTO |
| 31:02 | 7 | Edit | “extra breath before…” | YOURS |
Advanced Proof runs on our own dedicated AI hardware, in-house — your book is never handed to a third-party cloud service.
And it understands narration. Numbers, roman numerals and spoken variants are matched by meaning, so “forty” against “40” is never a false alarm — what gets flagged is what a human proofer would actually mark.
It replaces the listening, not the judgement. The engine does the hours of ear-work and brings you every suspect moment with the manuscript line beside it — you play it, decide, and only what you accept becomes a pickup. Nothing is re-recorded on a machine’s say-so.
The three things narrators actually get wrong: substitutions (a different word than the manuscript), omissions (words or lines skipped) and insertions (words added). Each catch carries its chapter, timestamp and a confidence score.
No. Numbers, roman numerals, spellings and pronunciation variants are matched by meaning, and every finding is confidence-gated — it was tuned on real audiobooks specifically to avoid wasting your time with formatting noise.
Your manuscript and your chapter audio — MP3, WAV, AIFF, M4A or FLAC. Everything stays in your private account area.
On ST Mastering’s own dedicated AI hardware, in-house. Your book is not shipped to a third-party cloud transcription service.
Four things: the reviewed findings, a numbered pickup script (chapter, time, reason, the exact sentence to re-read), a preview audio file with the original take of each pickup for tone-matching, and a downloadable report ordered by book position — timestamp and chapter in their own columns — that you can edit from directly.
AB ProofChk puts the manual proofing booth in your pocket — listen anywhere, drop timestamped notes with one thumb, walk away with a clean pickup report. Coming soon to the App Store.
Upload your book, review the catches, hand your narrator the pickup pack — same day.
Your first 20 minutes are proofed free — no card needed.