Description

The problem every AI-using author has

You paste a scene description into Claude or ChatGPT. You hit enter. You get back something that sounds nothing like you.

It’s smooth. It’s competent. It’s also generic AI prose — the contrast clauses, the abstract emotion-naming, the “It’s not just X, it’s Y” formulations, the closing summaries that read like a school essay.

You delete it. You try again. You spend twenty minutes hand-feeding the AI examples of your work, hoping it sticks. By chapter three, the AI has drifted back to its default voice and you’re scrubbing AI tells out of every paragraph.

The Author Voice Profiler fixes this once.

What it is

A structured voice pipeline that produces one document: a portable Voice Profile you own and reuse forever. Paste it into any AI writing tool, and the tool’s output sounds like you instead of generic AI.

Two modes

Mode A — Capture an existing voice. You have prose samples for the pen name. The plugin extracts the voice already on the page.

Mode B — Design a voice for a new pen name. You’re launching a brand new pen name with no writing yet. The plugin walks you through a directional questionnaire, calibrates against comp authors, and synthesizes a target voice you can refine before you write a single word.

How Mode A works

The capture pipeline runs in five steps:

1. You upload 2-3 prose samples from your existing work — actual scenes with narration, dialogue, and internal thought.
2. The plugin extracts your Voice DNA across nine observable dimensions: sentence rhythm, vocabulary register, narrative distance, dialogue style, punctuation fingerprint, internal monologue style, imagery patterns, pacing signature, and sentence structure. Every observation is concrete and anchored to your samples.
3. A short diagnostic questionnaire catches what the samples don’t show: your humor register, your banned words, the feeling you want at every chapter break, the cliches you actively work against.
4. Comp author calibration locks the voice neighborhood: you name 2-3 comp authors and the specific ways you diverge from each, plus one author you’d hate to be compared to.
5. Side-by-side validation test — the plugin writes a sample paragraph in the captured voice, you rate it 1-5, and the profile refines until it sounds like you.

How Mode B works

The design pipeline runs in six phases:

1. Pen name and market frame. Pen name (or placeholder), genre, subgenre, target reader, length tier, and the “why” behind the pen name.
2. Voice direction questionnaire. The standard diagnostic plus six Mode B-specific questions: imagined sentence rhythm, vocabulary register on a 1-5 scale, narrative distance, the first-page test, the dust-jacket test, and voices you actively don’t want.
3. Comp author anchoring. Comps are used as targets, not just neighbors — more directionally than in Mode A.
4. Voice synthesis. Builds a draft profile where every Voice DNA section is anchored to a comp element or a stated preference.
5. Demo paragraph and three-axis validation. A 200-300 word demo paragraph rated on Direction, Texture, and Differentiation. Refine and re-test up to three rounds.
6. Final delivery. Profile is marked as Mode B origin and includes a “first chapter checkpoint” instruction telling you to re-run in Mode A after you have 1-2 chapters of real prose.

What you get

One Voice Profile document containing:

• Voice Identity Statement — the headline of how your voice operates on the page
• Full Voice DNA — nine dimensions of concrete, observable patterns
• POV and tense defaults
• Tone calibration
• Comp author anchors and divergences
• Hard Rules (always do)
• Hard Avoid (never do — including specific banned words and AI tells)
• Voice Markers — short example phrases for the AI to pattern-match against
• Six paste-ready AI Platform Prompt Blocks for:
  • Claude (Project Instructions)
  • ChatGPT / Custom GPT (Custom Instructions)
  • Gemini Gem (Instructions field)
  • NovelCrafter (Codex / Style Guide)
  • Sudowrite (Style Description)
  • Generic (any chatbox)

Choose your delivery format

Run the plugin and pick one of three:

• Markdown in chat — paste-ready, lowest token cost, works in any text field
• Word document — formatted, downloadable, editable
• PDF — polished, ready to print or share

Who this is for

Indie fiction authors who:

• Use Claude, ChatGPT, Gemini, NovelCrafter, Sudowrite, or any other AI writing tool
• Are tired of scrubbing generic AI prose out of every output
• Have a voice they’re proud of and want to preserve when working with AI (Mode A)
• Are launching a new pen name and want a target voice locked in before they write a word (Mode B)
• Manage multiple pen names and need a separate, portable Voice Profile for each
• Want a portable, reusable profile instead of re-feeding samples every session
• Plan to use AI across multiple tools and want consistent voice everywhere

What this plugin does NOT do

• Does not rewrite your books
• Does not produce manuscripts, scenes, or chapters in your voice
• Does not critique the quality of your prose
• Does not recommend genre or market positioning

The Voice Profile is a tool. You still do the writing. This plugin makes sure the AI helps in your voice instead of fighting you.

Installation

Download the .zip file. In Claude, install via the plugin marketplace or drop into your plugins folder. Run /capture-voice or just say “capture my voice.”

Version 1.0.1