The field guide for private, local-first audiobook generation
Turning Fiction is designed to feel like a compact single-page app, not a bloated marketing site. This page keeps the home screen clean while documenting why the app works, how the workflow fits long-form reading, and where to find source material.
Why Turning Fiction
A focused tool for turning dense reading into portable listening.
The app is optimized for long-form text instead of one-off voice clips. Import a document, clean it up, choose a voice, and export audio without leaving the browser.
Local First / Privacy
Inference stays in-browser, so the app can remain private by default.
Document parsing, model loading, audio generation, and playback all run on your device. That reduces upload risk and makes the tool useful for drafts, notes, and personal reading material.
Why the Model Matters
A dedicated speech model behaves differently from a general text model.
Turning Fiction uses a specialized text-to-speech model for stable long-form narration, voice consistency, and better pacing across extended passages.
How It Works
A short workflow that matches the current app behavior.
- 1. Upload EPUB, PDF, DOCX, or TXT, or paste text directly.
- 2. Review the parsed text before generation.
- 3. Pick voice and quality settings, then generate audio.
- 4. Listen in-app, seek through synced text, and export.
Using Edit Mode Well
Edit mode is the control point for cleaner audiobook output.
After import, use Edit mode to remove headers, front matter, repeated page furniture, OCR artifacts, or passages you do not want read aloud. Once audio exists, toggle into the synced reading view with the book icon to follow along and scrub playback.
Formats and Limits
Best results come from clean source text and a quick review pass.
EPUB, DOCX, TXT, and pasted text usually produce the cleanest results. PDFs can work well, but scanned or messy layouts benefit most from a manual edit pass before generation.
Resources
A collection of community-recommended links and resources for finding digital books and reading materials.
Z-Library Subreddit/Wiki
Community-maintained guide and resources for finding digital books and alternatives.
Project Gutenberg
Public-domain ebooks in multiple formats for clean imports.
Standard Ebooks
Carefully proofed public-domain editions that are ideal for EPUB workflows.
Internet Archive
Public-domain collections and borrowable texts, depending on regional availability.
Best practice flow
A short routine that keeps exports cleaner and reduces reruns.
Use Upload when you already have a source document. Use Type when you want to paste excerpts, rewrite sections, or assemble a custom reading queue from multiple sources.
Before generating, skim the text in Edit mode and remove anything that sounds awkward when spoken aloud: copyright pages, chapter indexes, duplicated headers, footnotes, or OCR debris.
After generation, the synced reading mode is useful for checking pacing, jumping to a passage, or deciding whether a paragraph needs another edit pass.
What this page is for
A compact info surface that fits the product instead of competing with it.
The homepage stays minimal so the main action remains obvious: bring in text and start listening.
This route carries the explanatory weight, improves crawlable site surface, and gives users a place to understand the tool without disturbing the core app flow.