Audiobook repair & mastering

Master your complete audiobook without mastering audio engineering.

Checked against the requirements for ACX / Audible, Spotify / Findaway, Apple Books, Google Play, Kobo and Authors Republic

Upload your edited narration. ChapterMend repairs technical audio problems, matches every chapter and returns correctly formatted files with a complete plain-language report.

Secure upload coming soon
Private chapter uploads open with the founding beta. Nothing is uploaded from this page yet.
Join the founding beta
WAV, MP3 or FLAC · one file per chapter · up to 1 h 20 m each
Built from a 700+ audiobook production workflow We repair your recording — never regenerate it Your voice is never used to train models Files automatically deleted after delivery

You bring

  • Narration you already recorded and edited
  • Human or AI voiced — the repair works the same
  • One file per chapter, in any common format
  • Nothing else — no plugins, no settings

ChapterMend handles

  • Noise, hum and room-tone repair
  • Levels, peaks and noise floor to submission spec
  • Music intros and outros left as you mixed them
  • Spacing, MP3 formatting and the technical report
How it works

Three steps, and none of them are audio engineering.

1

Upload your edited chapters

WAV, MP3 or FLAC. One file per chapter.

2

ChapterMend diagnoses and repairs

It fixes safe technical problems and flags anything that should be re-recorded.

3

Download the complete package

Receive mastered chapters, organized delivery files and a plain-language report.

Before / after

See what ChapterMend detects and repairs.

Defects are located first, then repaired only where repairing them will not hurt the performance.

Gain-matched — louder never passes for better
0:000:42
Demo audio coming soon The waveform above is a diagram of the analysis, not a recording. Gain-matched before/after clips are being prepared.
The whole book, not one file

ChapterMend reads your book the way a producer does.

Every chapter is checked, repaired where it is safe, and honestly flagged where it is not — with exact timestamps for anything that needs re-recording. You always know where your book stands. The chapters below illustrate the format of that report.

Opening Credits
0:38
Passed
Chapter 1 Two clicks removed · levels normalized
32:10
Repaired
Chapter 2
28:44
Passed
Chapter 3 Heavy clipping at 14:22–15:04. Repairing this section would damage the narration — re-recording 42 seconds is the honest fix.
30:57
Re-record recommended
Closing Credits
0:52
Passed
Retail Sample First five minutes — worth a human listen before you submit
4:58
Manual review
Project-wide consistencyVolume matched across all 6 files
Example report

Every measurement, explained in plain language.

The same numbers a mastering engineer checks — with a verdict you can act on, not jargon. The figures below are an example of the report format, not a customer’s book.

RMS level
−21.4 dB
Passed
Within the −23 to −18 dB submission window.
Peaks
−3.6 dB
Passed
Nothing exceeds the −3 dB ceiling.
Noise floor
−68 dB
Repaired
Lowered from −54 dB without touching the voice.
Sample rate
44.1 kHz
Passed
Correct rate, constant throughout.
Bitrate
192 kbps CBR
Passed
Constant bitrate MP3, as retailers require.
Head / tail spacing
1.0 s / 3.2 s
Repaired
Room tone added so chapters breathe correctly.
Channels
Mono
Passed
Consistent across every file in the project.
Chapter duration
38 min
Warning
Long for a single chapter — fine to submit, worth a look.
A technical preflight. It is not a guarantee of retailer approval — human reviewers also judge performance, pacing and accuracy.
What ChapterMend is built to repair

The fixes engineers charge by the hour for.

You send a finished recording — narration, music beds and all. ChapterMend repairs and masters it. It never generates narration, and it leaves your mix where you put it.

Noise, levels, peaks, room tone and MP3 formatting run in the automated chain today. Click and pop repair, re-record detection and complete-book packaging are in testing — exactly what the founding beta is for.

In testing

Click & pop repair

Mouth clicks and plosive pops removed at the sample level, not smoothed over.

In the chain now

Hum reduction

Electrical hum and its harmonics pulled out of the recording, voice untouched.

In the chain now

Noise reduction

Hiss and room noise lowered conservatively — never the smeared “underwater” sound.

In the chain now

Consistent chapter volume

Chapter 12 sounds like Chapter 1, even if they were recorded weeks apart.

In the chain now

Room-tone correction

Natural silence at the head and tail of every chapter, taken from your own room.

In the chain now

Correct MP3 export

192 kbps CBR, 44.1 kHz, metadata stripped, encoded once from a lossless master.

In testing

Re-record detection

When processing would hurt the narration, we say so — with the exact timestamps to redo.

In testing

Complete-book packaging

Chapters named, ordered and zipped with credits and retail sample in place.

Built from a 700+ audiobook production workflow No voice generation — your recording comes back as your recording Every repair is listed, not hidden Independent — not affiliated with ACX, Audible or Amazon

Your audio is used to process your book and nothing else: it is never used to train models, and files are deleted automatically after delivery. The full data-handling and deletion policy will be published, and verified end to end, before the first customer file is accepted.

Founding beta

One book, one price, reviewed personally.

A small first group of books, each one checked by Joel before it goes back to you.

Founding beta
$49 one-time
Up to 8 finished hours
  • Complete-book technical diagnosis
  • Safe audio repair and mastering
  • Chapter-to-chapter consistency check
  • Plain-language technical report
  • Retail-ready file packaging
  • Founder quality review
  • Automatic deletion after delivery

Have a longer book? Contact us for a custom pilot quote.

Questions

What ChapterMend does, and what it doesn’t.

Does ChapterMend replace audiobook editing?

No. ChapterMend starts where your edit ends. You edit the performance — retakes, pacing, breaths you want gone — and ChapterMend handles the technical layer on top: repair, levels, spacing, formatting and verification across the whole book.

What problems can ChapterMend repair?

Loudness and peak levels, a noise floor that is too high, electrical hum, missing or wrong head and tail room tone, inconsistent volume between chapters recorded on different days, wrong sample rate or bitrate, and mixed mono/stereo files. Click and pop repair is in testing for the founding beta.

What problems require re-recording?

Heavy clipping, strong room echo, and misreads or missing text. Software can disguise these, but it damages the narration doing it — so ChapterMend flags them with exact timestamps and tells you to record a pickup instead of quietly making the audio worse.

Does ChapterMend guarantee retailer approval?

No, and be wary of anyone who does. ChapterMend verifies the measurable technical requirements. Retailers also run human review that judges performance, pacing and manuscript accuracy — things no mastering tool can pass on your behalf. ChapterMend is independent and not affiliated with ACX, Audible or Amazon.

What happens to my audio after processing?

It is used to process your book and nothing else. Files are deleted automatically after delivery, and the complete data-handling and deletion policy will be published — and verified end to end — before the first customer file is accepted.

Does ChapterMend generate or clone voices?

No. ChapterMend never generates speech and never clones a voice, and your audio is never used to train models. Whatever you send comes back as the same performance — repaired, levelled and correctly formatted.

Can I send AI-narrated audio?

Yes — if you already have the audio file, send it. ChapterMend measures and repairs audio; it neither knows nor cares how the recording was made, and AI-voiced files master exactly like any other. To be clear about what we are not: ChapterMend does not produce narration for you. There is no text-to-speech service here, no voice generation and no script-to-audio step — you bring a finished recording, we repair and master it. Where you submit it is your call: ACX and Audible do not accept AI or text-to-speech narration (Audible’s own Virtual Voice is the sanctioned route there), while Findaway/Spotify, Google Play and Apple Books each have digital-narration paths that permit it with disclosure. Check the platform’s current rules before you submit.

My chapters have music at the start and end. Is that a problem?

No. Intro and outro music beds are normal in audiobooks and ChapterMend masters the file as delivered — your music stays where you put it, at the balance you chose. Worth knowing: retailers measure the whole file, so a music bed that is much louder or quieter than the narration can pull the file’s overall levels around, and music running under a fade is not silence. If that affects your measurements, the report tells you which chapter and where, rather than quietly re-balancing your mix.

“Every rule ChapterMend enforces comes from audiobooks that actually shipped — including the painful lessons.”
Created by Joel Thielke using lessons from publishing more than 700 audiobooks across Audible and other platforms. He is personally reviewing every founding-beta book.