From Manuscript to Audiobook in 3 Steps

This is the practical how-to companion to our overview of writing and publishing in one place. If you want the dashboard walkthrough, you're in the right post.

Before you start

You'll need:

• A free MimicReader account (1 hour of audio generation included monthly)
• Your manuscript — as text you can paste, or you can type it directly
• About 30 minutes for a typical 60,000-word novel; 5-10 minutes for a short story

Step 1 · Write your chapters

Open Projects in your dashboard. Click New project. Fill in:

• Title — book title (you can rename later)
• Blurb — one-line description (used in the EPUB metadata and on the share page)
• Language — pick from 23 supported languages. This decides the narrator's pronunciation, so set it before generating audio.
• Genre + target audience — optional; helps the AI helpers if you use them

Once the project exists, you'll see a chapter list. Click New chapter for each chapter. The editor is plain markdown — paragraphs separated by blank lines, headings with ##, bold with **, italic with _. Everything saves continuously; no save button to click.

Pasting an existing manuscript? A few tips:

• Split chapters into separate chapter rows — easier to reorder and edit than one giant document
• Strip page numbers and running headers if pasting from a PDF — those become spoken words otherwise
• Smart quotes ("…") and em dashes (—) render correctly in TTS; ASCII (-) sometimes sounds like "minus"

Step 2 · Export EPUB (optional but recommended)

You don't strictly need an EPUB to generate the audiobook — MimicReader pulls chapters directly from the project. But exporting EPUB first is useful for two reasons:

1. You get a distributable ebook file you can upload to KDP, Apple Books, or send to beta readers
2. The EPUB process is the same code that the audiobook generator uses, so if EPUB looks right, the audiobook will too — it's a quick sanity check

In your project, click Export → pick EPUB. The file generates in a few seconds and downloads to your computer. Open it in Apple Books, Calibre, or any EPUB reader to verify chapter breaks and metadata look correct.

If something looks off (a chapter is in the wrong order, or a heading didn't render), fix it in the project editor and re-export. The fix takes seconds because the project is the source of truth.

What's inside the EPUB

• Cover image (if you generated one — see step 3 sidebar)
• Title page with title, author (your username), blurb
• Table of contents linking each chapter title
• One HTML file per chapter
• NCX/NAV navigation so it works on Kindle, Apple Books, and standard e-readers

Step 3 · Generate the audiobook

Open your project. Click Generate audiobook. You'll see voice and tier options:

Click Generate. The pipeline runs:

1. Concatenates your chapters into one long text
2. Chunks the text into TTS-sized segments (~200 chars per chunk by default)
3. Generates each chunk in parallel across 3 GPU workers (or 1 in cloud overflow mode)
4. Concatenates the audio with smooth pauses at sentence / paragraph / chapter boundaries
5. Normalizes to EBU R128 broadcast standard and encodes to M4A with chapter markers

For a 60,000-word novel, expect ~30 minutes of generation time (real-time factor ~0.3 with cached voice clone). You can navigate away — the job continues in the background and you'll get an email when it's done.

Out of free credit?

If your book is longer than the 1-hour free monthly credit covers, you have two options on the /pricing page:

• Subscribe £2.50/month — 2.5 credits monthly, non-cumulative, cancel anytime
• Buy a pack — £5 / £20 / £100 packs at £1 per credit, never expire

The two are fully independent — subscribers can buy packs on top, and pack-buyers don't need to subscribe.

Common questions

How long does generation take for a typical book?

Real-time factor with the default OmniVoice on our RTX 3090 is about 0.30 — meaning 1 hour of audio takes about 18 minutes of wall time. A 10-hour novel ≈ 3 hours of generation. You don't have to babysit it; the job emails you when done.

What if I want to change a chapter after generating?

Edit the chapter, click Generate again. The pipeline reuses cached chunks where it can, so partial edits (typo fixes) re-render fast. Big rewrites take roughly proportional time.

Can I download the audiobook to my phone?

Yes. The M4A is a standard audio file — works in Apple's Books app, VLC, Smart Audiobook Player on Android, any modern audio player. MimicReader also runs as a PWA with built-in player + offline mode if you want to keep everything in one place.

Can I sell the audiobook?

Yes. You own the audio you generate. Upload to ACX (Audible), Findaway Voices, Authors Direct, BookFunnel, or self-host. The £2.50/month subscription tier and PAYG packs both come with full commercial rights at ≥100h lifetime usage; for shorter projects, see our self-publishing guide.

Can my book be in multiple languages?

One project, one language per generation pass. If you want a bilingual edition (e.g. Polish + English versions), create two projects with the same text — set the language on each, generate separately.

Where to go next

• Why we built Projects: writing + audio in one platform
• Self-publishing your audiobook in 2026
• Already have an EPUB? Skip steps 1-2 and just generate audio