Nonfiction audiobooks. One person performs it, edits it, masters it, and answers for the files that ship.
I'm Clay. Twenty years as a live-event audio engineer, now narrating nonfiction.
Performance and production stay under one roof, so the person interpreting your manuscript is the same person accountable for what comes out of it.
Start with the samples. If the voice isn't right for your book, nothing else on this page matters.
Listen first
Sixty to ninety seconds each. Different demands, so you can hear how the read changes with the material.
LAUNCH BLOCKER: record six 60–150s samples into /samples/ before this page goes live. Narrate from books you own or public-domain text, and say which in the caption. A narrator site with no audio does not launch.
Then hear your own book
Within 48 hours you'll have a three-to-five minute excerpt of your own manuscript, narrated and mastered, plus a note on how the book should be produced and what it will cost.
Not a reel. Your pages, my read, so the decision is about your actual book.
One audition per book — a second read of the same pages would tell you the same thing. Your chapter is used only to record the excerpt and is deleted afterward on request.
By sending a chapter you confirm you hold the audio rights to it, or that you represent the person who does. You keep every right in your manuscript — sending it grants no licence beyond recording the excerpt, and I don't publish it, use it in marketing or show it to anyone without your written permission.
Want an NDA first? Ask, and I'll send one before you upload anything. Audition terms · How your files are handled.
Narration and post-production usually sit with different people. That works — most professional narrators either know production well or partner with an editor and engineer who do.
The version I offer is narrower: the person interpreting the manuscript is the same person who edits, masters and ships the files. Nothing gets handed across a gap, and there's no question about who owns a problem.
I've spent twenty years engineering speech under pressure — live rooms, difficult acoustics, an audience already seated, no second take. Gain staging, noise control, session continuity and file QC aren't skills I picked up for audiobooks.
That's what makes a fixed date and a correction guarantee safe to offer. It isn't what makes the read good — the samples above are the only evidence for that.
I read the book. Every name, every citation, every piece of jargon goes into a pronunciation list, and you approve that list before I record a syllable.
Four days. Longer if the bibliography is in German.
Fifteen minutes, recorded and mastered. You sign off on pacing, tone and pronunciation.
This is your exit. If the read is wrong for the book we stop here and you pay for the fifteen minutes.
The long part.
Punch-and-roll at the mic. Then the edit. Then a proof-listen against your manuscript, line by line.
You'll hear from me twice: once at the halfway mark, once when it's done. Otherwise I'm working.
Files mastered to spec and measured. The measurement report ships with them, so the numbers arrive at the same time the audio does.
Chapter-tagged files, opening and closing credits, a cut retail sample, upload-ready. I'll walk you through distribution if you want it, and stay out of the way if you don't.
Thirty days from checkpoint approval for a standard nonfiction book. The date goes in the contract, and the day counts above are the plan inside it.
Fixed project price, quoted from word count before you commit. Nonfiction runs about 9,300 words per finished hour, which is the planning estimate ACX itself publishes — your actual runtime will vary a little either way, and the quoted price doesn't.
| Narration and production | |
| With the Launch Kit | |
| Post-production only |
Quoted from $397 per estimated finished hour. Priced separately when the manuscript includes heavy tables, equations, extensive footnotes or citations read aloud, foreign-language passages, multiple speakers, or a pronunciation list beyond about 150 terms. Those are named in the quote, not discovered later.
Thirty days from checkpoint approval. Half up front, half on delivery.
Most people who add this are using the book as the front of a practice, not as a product. If that's not you, skip it.
Already recorded, or holding narrator files ACX won't accept. Edit, repair, proof-listen, master, retail sample, chapter delivery.
Run your master through the preflight tool below first, then send it over. If it isn't recoverable I'll say so instead of taking the work and failing slowly.
Capped at one a month.
Current availability
EDIT WEEKLY, BY HAND, HONESTLY: next production slot opens 2026-07-28.
If files are rejected or returned for a technical or production fault within my control — levels, noise, edits, room tone, file structure, credits, naming — I correct and resubmit at no charge until they're accepted.
This covers the audio. It doesn't cover metadata, cover art, rights disputes, content-policy decisions, author-supplied material, or a platform changing its requirements.
Thirty days from checkpoint approval. Every week past it takes 5% off the invoice, up to 20%.
The clock pauses for manuscript changes, late approvals, late pronunciation answers, added scope, illness or vocal injury, and anything outside either of our control. Those pauses are logged in writing when they happen, not argued about afterward.
You hear roughly fifteen minutes before full production starts. If the read isn't right for your book, we stop there and you pay for the checkpoint only.
You shouldn't have to commit to a full production to find out whether a voice can carry your book for nine hours. That's what the samples and the checkpoint are for.
Have a lawyer turn these into contract terms before you sign anyone. What's on this page is the summary, not the agreement.
Technical production
| Measurement | Target | What it's for |
|---|---|---|
| Peak | ≤ −3 dBFS | The headroom ACX requires for encoding |
| Speech RMS | −23 to −18 dBFS | Same perceived loudness, chapter to chapter |
| Noise floor | ≤ −60 dBFS | Silence that's actually silent |
| Delivery format | 192 kbps+ CBR MP3, 44.1 kHz | ACX's required submission format |
| Room tone | 0.5–1 s, head and tail | Chapters that don't start or end abruptly |
| Structure | Credits, retail sample, chapter tags | Opening and closing credits, a cut retail sample, correct file naming |
The measurement report ships with your files.
Human narration, and it stays that way.
No synthetic or cloned voice in any file I deliver, and no digital replica rights assumed without your separate written consent. That's a clause in the contract, not a marketing position.
Meeting these is the floor, not the point. A file can measure correctly and still be a poor listen — which is why the samples are at the top of this page and this section is near the bottom.
BEFORE PUBLISHING: confirm every figure against ACX's current Submission Requirements from inside your own account.
Free tool
Drop in an audio file for an estimate of where it sits against the ACX level targets. It's analyzed here in your browser — nothing is uploaded, nothing is stored, and I don't see it.
What this is and isn't. It's an unofficial estimate of four level and format measurements, useful for catching an obvious problem early. It is not affiliated with ACX and it does not predict whether a submission will be accepted.
ACX runs its own analysis on uploaded files and checks more than this does — and ACX says plainly that a file can pass automated analysis and still be rejected by a human for editing, extraneous noise or dynamics. Different tools also compute RMS differently, so treat any number here as approximate.
Validate final files with ACX's own audio analysis tool inside your project.
The legacy royalty model is being retired. The new model opened for enrollment on May 26, 2026 — 50% on exclusive distribution, 30% on non-exclusive. Titles on the legacy 40%/25% rates need to transition by the end of the year to stay distributed.
Exclusivity got much easier to exit. The distribution agreement still runs for a defined term, but a rights holder can now request a change of exclusivity or an early termination once the title has been on sale for at least 90 days. The practical seven-year lock is gone. The agreement structure isn't.
That second one matters. For years the most common reason authors skipped audio was getting stuck with one retailer for the better part of a decade. That's no longer the trade.
BEFORE PUBLISHING: verify both against ACX's current help documentation from inside your own account. Third-party guides still publish the old figures.
Audible publishes AI-narrated titles through its own Virtual Voice program, and there are authorized voice-replica programs as well. Those are legitimate routes. Unauthorized text-to-speech submitted as human narration is a different thing and ACX doesn't permit it.
If your book is a product and the budget is the constraint, Virtual Voice is a real option. Use it.
If the book is the reason someone should listen to you, the calculation changes. Synthetic narration reads accurately. It doesn't decide — it can't tell which sentence in your book is the one people will quote back to you, so it gives that sentence the same weight as the one before it. Emphasis is a judgment about meaning.
Listen to the samples, then decide. That's a better basis than anything I could argue here.
Separate service
For coaches, founders, consultants and speakers, the voice is often part of the product. Your readers have never heard you, and for some books that's the whole opportunity.
Narrating your own book is harder than it sounds and most authors who try it alone give up around hour three. What makes it work is direction — someone running the session, catching the reads that drift, and handling everything after the microphone.
That's a different engagement: performance coaching, directed sessions, recording, pickups, edit, proof, master, and delivery prep. Same production standard, your voice.
LAUNCH BLOCKER: replace with real testimonials and a credits list as you earn them — specific result, timeframe, a number where one exists. Fabricate nothing. Until then, publish the paragraph below and nothing else.
There's no credits list here yet. I'm building one, and the rate reflects that.
Arithmetic, not a promotion. No deadline attached, and it isn't changing this month.
Nonfiction runs about 9,300 words per finished hour, which is the planning figure ACX publishes. A 79,000-word book estimates to roughly 8.5 hours. Actual runtime varies with the material and the read — the quoted project price doesn't.
Thirty days from checkpoint approval for a standard nonfiction book, with the clock pausing for manuscript changes and late approvals. The date goes in the contract.
Listen to the six samples first. If they're close, send a chapter and hear your own pages. If you've already engaged, the 15-minute checkpoint is the exit — you approve the read before full production and pay for the checkpoint only if you stop.
No. The agreement runs for a defined term, but you can request a change of exclusivity or an early termination once the title has been on sale 90 days. We work through both options against what you want the book to do before you decide.
Sometimes that's the right call and I'd rather you make it knowingly. The difference at this price is that performance, editing, proofing and mastering sit with one person, the date is contractual, and production faults get corrected at no charge. At the low end that varies a lot — though plenty of those projects go fine.
If it isn't final, it isn't ready. Change a sentence after recording and that chapter gets re-recorded, re-edited and re-mastered, and it's billable. Finish the book first. There's no rush on my end.
No. Nonfiction only. Fiction is a different craft and I'd be worse at your book than the specialist I'd refer you to.
Narrator errors — misreads, mispronunciations against the approved list, audio faults — are corrected at no charge, always. Changes to an already-approved performance direction are limited and quoted. Manuscript changes are billable, because they mean re-recording.
Nothing leaves your machine. The analysis runs in your browser. Nothing is uploaded, nothing is stored.
Yes. Downloadable samples, studio specification and a credits list are on the producers page, and I'm available for remote-directed sessions.
If the voice is close, send a chapter and hear your own pages inside 48 hours.