[Tutorial] Using Spoken.Press to Create a Nonfiction Audiobook with AI

I just used Spoken.press to create a nonfiction audiobook with AI, using my own AI-cloned voice—and the process was surprisingly easy. So here I am to share it with you!

If you want a visual, here’s my video walkthrough on YouTube.

If you’ve written a nonfiction lead magnet to support your coaching or consulting business, it’s got a job to do: grow your business, increase your visibility, and help your ideal clients trust you faster.

But what if your ideal clients are audio-first learners? Creating an audio version of your nonfiction book or lead magnet used to be fairly cost-prohibitive, or require you to book a studio and spend hours recording. Spoken.press makes that so much easier.

In this tutorial, I’ll walk you through how I used Spoken.press to turn my nonfiction book (The 10-Step Impact Book Blueprint) into an audiobook, using a cloned version of my own voice.

This process is especially useful if:

• You’re a coach, consultant, or expert with a short authority book or lead magnet
• You want an accessible audio version for readers who can’t or don’t like to read
• You don’t want to spend thousands on a pro narrator (yet)

I’ll show you each step, plus a few things I really wish I’d known before I hit “Publish.”

First—I’m not an affiliate of Spoken. I am friends with the founder and have been on the advisory committee for the past year as they fine-tune their product for authors. I use and like their product—and hope you will, too!

What is Spoken.press (and When Is It a Good Fit)?

Spoken is an AI audiobook studio. You upload your manuscript, choose a voice from their selection of real narrators (who license their voices through ElevenLabs), AI voices, or even a cloned version of your own, and Spoken generates a full narration for you.

You can then proof and tweak individual passages before you pay to publish.

I’ve narrated my short fiction with Spoken in the past, using ElevenLabs voices because I like the idea that a real human narrator is getting paid for their work. For nonfiction, Spoken is great for:

• Short authority books and lead magnets
• Coaching frameworks you want clients to listen to on the go
• Accessibility-friendly versions of your existing book or workbook

If you’re creating a full-length commercial audiobook that you plan to sell widely on Audible and other platforms, you may still want to hire a professional narrator, or narrate it yourself.

But for a 5–20k word coaching book or lead magnet that needs to be useful and accessible, Spoken is a fantastic way to create a nonfiction audiobook with AI.

The pricing is $20 per finished 5,000 words for non-subscribers, and $10/5k words for subscribers. (Subscription is $50/month.) I recently paid $120 for someone on Fiverr for someone to record a 2-minute cinematic reel, for comparison.

—> Further reading: ROI of a Business Book (Case Studies & Calculator)

Step 1: Prep Your Manuscript for Audio

Before you ever log into Spoken, grab your manuscript and do a quick pass with audio in mind. You can absolutely tweak things later, but prepping now saves time.

(Ask me how I know.)

1. Start with a final draft: Typos, weird formatting, and placeholders (“TK example here”) will come through in your audio.
2. Strip out formatting that doesn’t matter for sound: While Spoken allows you to upload a PDF, it’s better to copy-paste the raw text rather than just popping in your PDF lead magnet. Spoken will include your page numbers, “copyright 2025 Jessie Kwak” on every page, etc. Clean, copy-pastable text is best.
3. Flag anything that might be awkward out loud.
   • Arrow flows like Book → Website → Course
   • Fill-in-the-blank workbook prompts
   • Visual diagrams or frameworks

We’ll talk more about how to rewrite those for audio in a minute.

Step 2: Set Up Your Account and Clone Your Voice

From the Spoken home page, click into the Studio. That’s where the magic happens.

To clone your voice:

1. Go to Voices → Manage my personal voice.
2. Spoken will give you a short script to read.
3. Use your best microphone. I used a Blue Yeti and got as close as I could in a quiet room. The file I submitted still sounded a bit echoey, but in the final audiobook it sounded pretty good.
4. Read the script at a normal pace, like you’re talking to a client.

Spoken uses this sample to build a custom voice that sounds like you.

When I listened back, it sounded like “me layered over a default female narrator” — slightly more sing-song than I would naturally read, but absolutely good enough for a lead magnet and accessibility version.

If you don’t want to use your own voice to create a nonfiction audiobook with AI, you can skip this step and choose one of Spoken’s built-in voices later. But it seriously took five minutes to set up, and I think nonfiction readers do want to hear the author’s voice.

Step 3: Create a New Project and Import Your Book

Once your voice is ready, it’s time to feed your book to Spoken.

1. In the Studio, click Create → New project.
2. Give your project a name (the title of your book).
3. Choose to paste in your text manually.
4. Paste your entire manuscript into the text box.
5. Click Confabulate.

Spoken will:

• Detect the language
• Figure out if it’s fiction or nonfiction
• Guess at the category (it correctly tagged mine as business / careers / investing)
• Run a quick safety/content analysis

This doesn’t change your content; it just helps Spoken optimize the narration behind the scenes.

Step 4: Break Your Book into Passages

Next, click Make passages.

Spoken will automatically break your book into smaller chunks, or “passages.” This makes it easier to edit specific sections and regenerate just one passage if something sounds off.

You’ll see a list of passages down the screen.

Step 5: Assign a Voice (Your Clone or a Library Voice)

Now it’s time to choose who reads your book. Click Assign voice, and you’ll see a few options:

• Library voices. These include AI voices and licensed 11Labs voices (real actors).
• Custom voices. This is where your cloned voice lives.

Select your preferred voice and click Preview. Spoken will read a short passage so you can get a feel for pacing, tone, and clarity.

If you’re using your cloned voice and it sounds a little more dramatic or melodic than you usually speak, that’s normal. Spoken is interpolating from your sample. For my nonfiction book, I decided it was totally fine — especially given how quickly I could get a full audiobook out of it.

(You can see me react to my cloned voice in real time in the YouTube video tutorial.

Once you’re happy, click Select this voice to apply it to the whole project.

Step 6: Generate the Full Audiobook

When your passages are ready and your voice is assigned, click Make Spoken.

Spoken will work its way through each passage, generating audio for your entire book. You’ll see the progress happen as the little icon beside the passages turned green.

For my 13,000-word nonfiction book, this took around ten minutes—which, compare that to recording the whole thing myself: I’d still be in chapter two and my voice would be wrecked for the day.

When Spoken finishes, the button at the bottom will change from “Updating” to Publish. You could publish now, but please don’t.

First, you need to prooflisten.

(Is that a word? It is now.)

Step 7: Prooflisten and Fix Glitches (Do Not Skip This)

Even when the overall quality is great, AI narration is not perfect. This is where your human ears come in. On my first pass, most of the audio sounded really good. But I still found:

• Mispronounced words (especially jargon, names, or less common terms)
• A couple of passages where my cloned voice randomly shifted into an Australian accent, lol
• Slightly odd phrasing choices that were fine on the page but clunky out loud

If I had just trusted the AI and hit “Publish,” I would never have caught those.

How to use forced AI Audio re-narration

The good news is that Spoken makes it easy to fix issues.

When you find a problem, just select the passage, adjust the text if needed (for spelling or clarity), and use Spoken’s option to force re-narration for that passage. Or, if it’s an accent problem, just try hitting force re-narration and see if it gets fixed on its own.

Sometimes I found it necessary to use a phonetic spelling to force the robot to read the right word. For example, “lead magnet” —> “leed magnet.”

Is it slightly tedious? Sure. But it’s still dramatically faster than recording your entire audiobook. I was mostly cleaning my office and sketching on my iPad as I listened, and I’d just hop over to my computer and fix mistakes whenever I heard them.

Here are three non-fiction specific fixes you’ll probably have to think about when you create a nonfiction audiobook with AI:

1. Rewrite arrow flows and shorthand

Nonfiction books often have passages that are great on the page, but useless when read verbatim.

For example, I had lines like:

Book → Website visit → Course purchase

On the page, that’s clear in context. In audio, it sounds like your narrator is reading out math.

I rewrote the passage as:

“When someone buys your book, that should lead them to your website, which then guides them to purchase your course.”

General rule: if something is visual shorthand, rewrite it as a full sentence your listener could follow while driving.

2. Fix fill-in-the-blank workbook language

My book also included workbook-style prompts like:

“My goal is to ______.”

For the audiobook, I changed lines like that to:

“Your goal is… blank. Say it out loud or jot it down in your notebook.”

You can also try:

• “Take a moment to finish this sentence: ‘My goal is…’”
• “Pause the audio and write down your answer.”

Speak directly to the listener and give them a clear instruction.

3. Add audio descriptions for diagrams and visuals

If your book includes diagrams, Venn circles, models, flowcharts, etc., none of that automatically translates to audio. Describing them is essential for accessibility. You’ll need to narrate it in a way that makes sense by ear.

For example:

“Imagine three overlapping circles. One is labeled ‘Your expertise,’ one is ‘Your audience’s pain points,’ and one is ‘Your business goals.’ The sweet spot where all three overlap is where your best book ideas live.”

You can add these descriptions directly into your text in Spoken, then regenerate those passages so the narration includes them.

A quick note: If you’re selling your audiobook alongside an ebook via Amazon’s KDP/Audible’s ACX, you can’t change the text too much without breaking WhisperSync. So be aware of that.

For more on how to make your online content more accessible and reach a bigger audience, check out Content for Everyone by Jeff Adams and Michele Lucchini.

Step 9: Publish and Pay

Once you’ve proofed your audiobook, you’re ready to publish. Just upload a cover image (2400×2400 pixels), adjust your description, and hit Publish. You can choose to publish it to their platform, or download your files and take them wherever you like.

When I created my 13,000-word nonfiction audiobook, it estimated around $60 to publish. You can edit and regenerate as much as you like before that point; you only pay once you’re happy with the final product.

(Obviously, pricing may change over time, but the model is designed to let you experiment before committing.)

Here’s the final result!

Overall, using Spoken.press to create a nonfiction audiobook with AI, using my own AI-cloned voice was really simple. It took me about an hour—and then I spent the rest of the day making tutorials about it.

If you try it, I’d love to hear your feedback!