Write a Book with Voice Dictation and Obsidian

The Author Who Writes While Hiking

Kevin J. Anderson has written over 180 books - sixteen million words - and most of them while hiking through the Rocky Mountains.

"I've got myself so trained that I can't sit there and stare at the screen and be very creative," Anderson explains in a recent interview. "I have to be out walking and moving around. I go hiking and mountain climbing, and everything's wonderful."

His workflow: outline chapters in advance, take notes for two or three chapters, hike a trail while dictating into a recorder, then send the audio files for transcription. He walks outbound until finishing a chapter, which means he has exactly enough time to dictate another chapter heading back to the trailhead. A full day of hiking produces around 10,000 words.

This isn't a quirky exception. It's a glimpse at what writing looks like when you remove the keyboard bottleneck and let storytelling return to its oral roots.

The technology has finally caught up to make this workflow accessible to everyone. Modern AI transcription eliminates the expensive typing services Anderson used for decades. Tools like Obsidian provide the organizational backbone. And voice-to-text tools like AICHE go beyond raw transcription - they polish your speech into clean, publication-ready prose automatically. No more hours fixing punctuation and cleaning up filler words.

Research increasingly shows that walking while creating isn't just pleasant - it's cognitively superior.

The Science of Walking and Writing

A Stanford study by Oppezzo and Schwartz found that walking increased creative output by 81% compared to sitting. Not a small boost. A near-doubling of creative ideation.

The study tested 176 participants across four experiments. People walking on a treadmill facing a blank wall produced twice as many creative responses as people sitting. Walking outdoors performed similarly. The creative benefit persisted even after sitting back down - a residual effect that continued to boost ideation after the walk ended.

"We're not saying walking can turn you into Michelangelo," lead researcher Marily Oppezzo noted. "But it could help you at the beginning stages of creativity."

For writers, this maps perfectly onto first-draft production. The generative phase of writing - getting ideas out, following narrative threads, capturing dialogue as it forms in your mind - benefits most from this creativity boost. Editing and revision, which require focused analytical thinking, can happen at the desk later.

The physical benefits compound the creative ones. Writers who spend 6-8 hours daily at keyboards face real risk of repetitive strain injury. Voice dictation eliminates the repetitive hand motions that cause RSI, carpal tunnel, and chronic wrist pain. Some professional authors credit dictation with saving their careers after injury made typing painful.

Speech also transmits information faster than fingers. Research published in Science Advances found human speech averages 39 bits per second across all languages. Typing during composition - when you're creating original content rather than transcribing - averages around 6 bits per second. That's a 6x speed difference before accounting for the cognitive overhead typing demands.

The Modern Dictation-to-Draft Workflow

The traditional dictation workflow required expensive transcription services or hours of self-transcription. AI changed this completely.

Modern tools can transcribe speech with 98-99% accuracy. But raw transcription still leaves you with a wall of unpunctuated text full of "um" and "uh." That's where AICHE differs from basic transcription tools.

AICHE's AI enhancement layer transforms speech into polished prose automatically. You speak naturally - pauses, filler words, run-on thoughts and all. The output arrives with proper punctuation, cleaned grammar, and structured paragraphs. The "unstructured rambling" objection to voice input becomes irrelevant when AI handles the structure.

This matters enormously for authors. You don't want to spend an hour cleaning up transcription for every hour of dictation. AICHE's "Polish" mode produces first-draft-quality prose directly from speech, cutting the post-processing time that made dictation impractical for many writers.

Here's a practical workflow combining dictation with Obsidian:

Before the session:

Outline your chapter in Obsidian with major beats
Review your character notes and any continuity details
Take physical notes for 2-3 chapters if going on a long walk

During dictation:

Speak naturally without worrying about punctuation
State headings aloud: "Chapter 12" or "Scene: The Confrontation"
Keep moving - walking maintains the creative flow
Don't stop to correct mistakes; mark them with a verbal note ("fix that") and continue

After the session:

Upload audio for AI transcription
Paste the transcribed text into Obsidian as a working document
Use Obsidian's Note Refactor plugin to split by headings into chapter files
Brief review for major errors, then move on - don't edit during the drafting phase

Setting Up Obsidian for Book Projects

Obsidian's folder-based structure maps naturally onto book organization. A dedicated vault for each book keeps everything contained.

Recommended folder structure:

My-Novel/
├── 00-Prep/
│   ├── outline.md
│   ├── characters.md
│   └── worldbuilding.md
├── 01-Drafts/
│   ├── 01-chapter-one.md
│   ├── 02-chapter-two.md
│   └── ...
├── 02-Research/
│   └── reference-notes.md
└── 03-Marketing/
    └── blurb-ideas.md

Use consistent naming with number prefixes to keep files in order. This makes reorganization simple - renaming 03-chapter-three.md to 05-chapter-five.md moves it in the sequence without copying content.

Essential plugins:

The Longform plugin transforms Obsidian into a proper manuscript manager. It tracks word counts per scene and project, compiles separate chapter files into a single manuscript for export, and supports multiple draft versions within one vault. One author has written over 70 books using Longform as their core workflow tool.

Note Refactor handles the dictation-to-chapters split. After dumping a long transcription into Obsidian, it can automatically divide the content into separate files based on your spoken headings.

The Kanban plugin works well for outlining. Create a board with columns for Act 1, Act 2, Act 3 - or whatever structure fits your story. Each card represents a scene or chapter. Drag them to rearrange plot points before committing to prose.

Dataview enables task tracking across your manuscript. Add checkboxes for issues noticed during dictation: "- [ ] Fix continuity about hero's car in Chapter 3". A Dataview query in a master To-Do note can aggregate every unchecked box from your draft files into one list.

The Obsidian + AICHE Integration

This is where the workflow becomes seamless. AICHE Voice for Obsidian is a native plugin that brings polished dictation directly into your vault.

Install in 30 seconds:
Settings → Community Plugins → Browse → Search "AICHE Voice" → Install → Enable → Sign In

How it works:

Click the mic icon in Obsidian's left ribbon (or use a hotkey)
Speak your scene naturally - the status bar shows recording time
Click to stop
Polished text appears at your cursor, ready for your manuscript

Why this matters for book writing:

The "Polish text" toggle transforms raw speech into clean prose. Whisper-based plugins dump exactly what you said - every "um," every false start. AICHE's AI removes filler words, fixes grammar, adds punctuation, and structures sentences. You speak like you're telling a friend about your scene; you get prose that reads like you typed it carefully.

Features designed for long-form work:

Voice Notes folder: All dictation goes to a configurable folder (default: "Voice Notes"). Keep your manuscript chapters separate from dictation dumps.
Category organization: Toggle auto-sorting into Work/Life/Ideas/Misc subfolders. Useful for separating novel dictation from research notes.
Audio attachment: Optionally embed the .webm recording with your note. Some authors keep the audio for reference when editing - hearing how you said a line helps with rhythm.
Offline queue: No wifi on the trail? Recordings save locally and process automatically when you're back online. Nothing gets lost.
New Voice Note command: Creates a new timestamped note and starts recording immediately. One action to capture a sudden idea.

For the walking writer: The plugin works on iOS and Android through Obsidian's mobile apps. You're hiking, an idea strikes, you pull out your phone - and the text goes directly into your novel's vault. No separate app, no syncing later, no copy-paste.

For desk sessions: The AICHE desktop app (available at aiche.app/profile?section=downloads) adds a global hotkey (Ctrl+Alt+R on Windows/Linux, Control+Option+R on Mac) that works in any application. Research in your browser, dictate notes. Writing in Scrivener, dictate prose. The same $4.99/month subscription covers both the Obsidian plugin and desktop app.

Privacy for manuscripts: Audio is purged within seconds of processing. Zero retention. For authors protective of unpublished work, this matters. Your novel-in-progress never sits on someone else's server.

Markdown for Manuscripts

Markdown's simplicity makes it ideal for long-form writing. No formatting dialogs. No wrestling with styles. Just text with lightweight markup that stays out of your way.

Each chapter file uses basic structure:

# Chapter 12: The Revelation

Sarah pressed her back against the cold stone wall...

## Scene Break

The guards found nothing. They wouldn't.

Headings (#, ##) create hierarchy for navigation and later compilation. When you export to Word, PDF, or EPUB using Pandoc, these headings become proper chapter titles and scene breaks.

YAML frontmatter at the top of your main file captures metadata:

---
title: "The Gathering Storm"
author: "Your Name"
date: 2026-01-29
status: first-draft
wordcount-target: 80000
---

This metadata can feed into Pandoc for generating title pages and professional formatting during export. It also helps track manuscript status across multiple projects.

The Psychology of Separating Creation from Editing

The single most important principle: dictate first, edit later.

Your brain handles creation and criticism differently. Trying to do both simultaneously creates cognitive conflict - the inner editor interrupting the inner storyteller. Dictation naturally enforces this separation because you can't easily backspace and revise while speaking.

Anderson's hiking workflow exemplifies this. He can't stop mid-trail to edit a paragraph. He speaks forward, captures the story, and deals with problems later. The physical movement prevents the stopping-to-tinker that derails so many writers.

This separation applies to the full session. Don't review transcription until after dictating. Don't fix typos during the creative phase. Don't reread yesterday's chapter before writing today's. Move forward. The editing phase exists specifically for fixing problems; the drafting phase exists specifically for creating them.

If you notice an issue while dictating, mark it verbally and continue: "Note to self: check if the timeline works here. Anyway, Sarah walked into the..." Then keep going. Those verbal notes become visible markers in the transcription that you can address during editing passes.

Building Your Dictation Practice

Dictation is a skill. Expect awkwardness at first.

Start with low-pressure exercises:

Read your own past writing aloud to hear your voice speaking your prose
Answer questions about your story's plot or characters (have ChatGPT generate 20 questions about your premise)
Describe a scene you've already visualized but haven't written
Dictate your daily journal or emails before tackling fiction

Voice care matters. Build up gradually from 30 minutes to a maximum of 2-3 hours per session. Stay hydrated. Speak naturally rather than adopting an artificial quiet or monotone voice - both strain your vocal cords. Take breaks.

The rhythm develops over time. Early sessions may produce choppy, disconnected prose. After a few weeks, you'll find yourself speaking in paragraphs, naturally pausing at paragraph breaks, and maintaining narrative flow across longer stretches.

What This Workflow Actually Produces

Writers who fully adapt to dictation routinely report 2-3x their previous first-draft output. The combination of speaking speed (160+ WPM versus 40 WPM typing) and reduced fatigue compounds into dramatic productivity gains.

But speed isn't the only benefit. Many authors find their dictated prose has a more natural rhythm - particularly dialogue. Speaking your characters' lines aloud catches clunky phrasing that looks fine on screen. The prose sounds better because it was literally spoken.

The health benefits matter for career longevity. Writers in their 40s and 50s who've accumulated decades of typing strain often credit dictation with extending their working lives. Removing the physical bottleneck means you can write as long as you can think.

And the walking component adds something harder to measure: thinking time. The ideas that emerge during a hike feel different from ideas that emerge while staring at a cursor. They've had room to develop, connections to form, problems to solve themselves without conscious forcing.

Try This Tomorrow

Pick one scene from your current project - ideally one you've outlined but haven't drafted. Go for a 20-minute walk with your phone. Open AICHE in Obsidian (or the desktop app). Speak the scene without stopping to correct anything.

When you return, review what you captured. With AICHE's polish mode, it won't be the messy transcription you'd get from basic dictation - it'll be readable first-draft prose. And it will exist, which is more than most unwritten scenes can claim.

The discomfort is normal. Your brain expects typing. Speaking prose feels foreign because you've trained for years on the keyboard pathway.

Give it three sessions before judging. By the third walk, something shifts. The words come faster. The self-consciousness fades. You start thinking in spoken paragraphs rather than translating typed ones.

Kevin J. Anderson has written sixteen million words this way. The technology that required expensive transcription services for him is now available through AI-powered tools like AICHE. The organizational systems that required custom software now exist in free tools like Obsidian.

The workflow is proven. The tools are ready. Install AICHE's Obsidian plugin, go for a walk, and speak your next chapter into existence.