Skip to content
✍️

Writing Style Guide

A style guide your AI reads before it drafts. Voice, banned words and habits, patterns, real examples. The draft starts closer to your voice, not generic AI default.

Requires an AI connected to your Hjarni account via MCP.

https://hjarni.com/templates/writing-style-guide

Copy this URL and paste it into Claude or ChatGPT to install the template.

How to use

  1. 1 Share this page. Paste this URL into Claude or ChatGPT. Your AI reads the template definition and installs it.
  2. 2 Folders, tags, and instructions appear. Your AI creates the full structure in your Hjarni account, ready to use.
  3. 3 Start adding notes. The AI instructions guide your AI on where to put things and how to organize them.

A style guide your AI reads before it drafts. The draft starts closer to your voice, not generic AI default.

A style guide pasted into one prompt helps for one chat. A style guide saved in Hjarni helps every chat. Every draft. Every reviewer.

Before and after

Same prompt. Same model. The only difference is the style guide.

Generic AI draft

"Our revolutionary new pricing model is designed to seamlessly unlock unprecedented value, empowering teams to supercharge their workflows while leveraging best-in-class features."

After the AI reads the style guide

"New pricing. Simpler tiers, same features, no surprises. Teams pay per seat."

The AI did not get smarter. It read your rules first.

Two ways to use it

Draft mode

Ask Claude or ChatGPT to draft something. The AI reads the Voice, Banned Words and Habits, and Patterns folders before writing. The draft starts closer to your voice. You edit less.

Review mode

Paste an existing draft and ask Claude or ChatGPT to review it against the style guide. It flags banned words, tone drift, unsupported claims, and places where the structure does not match your patterns. The AI becomes a first-pass editor that knows your rules.

Starter notes included

  • Voice Principles. A short list of rule-shaped principles.
  • Banned Words and Habits. Pre-loaded with the most-flagged AI tells (em dashes, "supercharge", "leverage", "seamless").
  • Reference Example. A short product update in the voice.
  • Pattern: Open with the news. One opening pattern with example.
  • Review checklist. What to flag when reviewing a draft.

Replace each with your own material once installed.

A workflow that earns the template's keep

You ask Claude: "Draft a tweet announcing our new pricing page." Claude reads the Voice folder, the Banned Words and Habits folder, and the Patterns folder, then drafts in your voice. You edit one word. You ship it.

A new writer joins the team. They connect their Claude. Their drafts start sounding like the team's voice on day one. The style guide did the onboarding.

Related pages

Structure

Tags

voice banned pattern example

Folders

Writing Style Guide
The voice your team writes in. Claude and ChatGPT read it before drafting or reviewing.
Has AI instructions
5 starter notes
Voice
The principles. Direct. Opinionated. Short sentences. Talk to one person.
Has AI instructions
Banned Words and Habits
Words, phrases, punctuation, and structures the brand avoids.
Has AI instructions
Reference Examples
Real published pieces in your voice. The source of truth for tone.
Has AI instructions
Patterns
Recurring structural patterns. How you open, how you close, how you handle lists.
Has AI instructions

For your AI

Share this page with your AI. It reads the definition below, creates the folders, instructions, tags, and starter notes in your account.

Show template definition

Install steps for AI agents:

  1. Check existing tags with tags-list. Only create missing ones with tags-create.
  2. Create containers top-down using containers-create, noting the returned IDs. Use parent_id to build the hierarchy.
  3. For each container with llm_instructions, call instructions-update with level: "container" and the container's ID.
  4. Create any seed notes using notes-create, placing them in the correct container by ID. Use container_path to resolve which container.
  5. Discuss any customizations with the user before or after installing.
---
name: Writing Style Guide
description: 'A style guide your AI reads before it drafts. Voice, banned words and
  habits, patterns, real examples. The draft starts closer to your voice, not generic
  AI default.

  '
tags:
- voice
- banned
- pattern
- example
containers:
- name: Writing Style Guide
  description: The voice your team writes in. Claude and ChatGPT read it before drafting
    or reviewing.
  llm_instructions: |
    This is a writing style guide. The goal is to get drafts closer to the team's voice on the first pass and to make review mode a real first-pass editor.
    - Before drafting any external copy, read Voice, Banned words and habits, and Patterns. Match what is there.
    - Never use any word, phrase, punctuation, or structure listed in the Banned words and habits folder. If a draft contains one, flag it and rewrite.
    - When reviewing an existing draft, flag banned words, tone drift, unsupported claims, and places where the structure does not match the patterns.
    - Prefer real Reference examples over invented imitations. When in doubt about tone, sample three recent examples and match them.
    - Do not invent new structural patterns unless the user asks for one.
  children:
  - name: Voice
    description: The principles. Direct. Opinionated. Short sentences. Talk to one
      person.
    llm_instructions: |
      Use this folder for principles, not examples.
      - When asked to draft anything in our voice, read this folder first. Match the rules exactly.
      - Keep notes short and rule-shaped: one principle per note when it earns the space.
      - When the user describes a new principle in conversation, suggest capturing it here.
  - name: Banned Words and Habits
    description: Words, phrases, punctuation, and structures the brand avoids.
    llm_instructions: |
      Use this folder for things to remove from drafts.
      - Never use any word, phrase, punctuation, or structure listed in this folder. If a draft contains one, flag it and rewrite.
      - The list covers more than vocabulary: it can include punctuation habits (em dashes, en dashes), fake contrast ("not X, but Y"), and over-polished AI tics.
      - When you spot a new tell in user-supplied drafts that the brand wants to ban, suggest adding it here.
  - name: Reference Examples
    description: Real published pieces in your voice. The source of truth for tone.
    llm_instructions: |
      Use this folder for source-of-truth examples in the team's actual voice.
      - When in doubt about tone, sample three recent examples from this folder and match them.
      - Prefer real examples over invented imitations.
      - Never rewrite an example to polish it. The point is to preserve the voice as it was shipped.
  - name: Patterns
    description: Recurring structural patterns. How you open, how you close, how you
      handle lists.
    llm_instructions: |
      Use this folder for structural recipes.
      - Use these patterns when drafting. Do not invent new ones unless the user asks.
      - One pattern per note: name the pattern, give a short description, link to one or two Reference Examples that use it.
  notes:
  - title: Voice Principles
    body: |
      A starter set. Replace with your real principles.

      - Direct. State the point first.
      - Opinionated. Take a side. Then defend it briefly.
      - Short sentences. Long ones earn their length.
      - Talk to one person. Use "you", not "users".
      - Concrete examples beat abstract claims.
      - Avoid hedge words. "Probably" and "kind of" rarely add information.

      This is a starter note. Replace it with your real voice principles.
    tags:
    - voice
    container_path: Writing Style Guide > Voice
  - title: Banned Words and Habits
    body: |
      A starter list. Replace with the words and habits your brand actually avoids.

      ## Banned words and phrases
      - revolutionary
      - game-changing
      - supercharge
      - leverage
      - seamless
      - robust
      - unlock
      - cutting-edge
      - empower
      - best-in-class
      - synergy

      ## Banned punctuation
      - em dashes
      - en dashes

      ## Banned structures
      - "Not X, but Y" fake contrast.
      - Three-adjective stacks ("powerful, intuitive, beautiful").
      - Closing with a rhetorical question.

      This is a starter note. Replace it with your real banned list.
    tags:
    - banned
    container_path: Writing Style Guide > Banned Words and Habits
  - title: 'Reference Example: Short product update'
    body: |
      *Replace this with a real published piece you wrote in your voice.*

      A short illustrative example of the shape:

      > New pricing. Simpler tiers, same features, no surprises. Teams pay per seat. Solo plans keep the free tier.

      Notes on what makes this fit the voice:
      - Direct first sentence.
      - No buzzwords.
      - No fake contrast.
      - One paragraph, three lines.

      This is a starter note. Replace it with a real piece in your voice.
    tags:
    - example
    container_path: Writing Style Guide > Reference Examples
  - title: 'Pattern: Open with the news, not the warm-up'
    body: |
      State the news in the first sentence. The rest of the piece earns the reader's time.

      ## When to use
      Product updates, announcement posts, change logs.

      ## Example
      - Good: "New pricing. Simpler tiers, same features."
      - Avoid: "We are thrilled to announce that today marks an exciting milestone for our team."

      ## Reference
      - [[Reference Example: Short product update]]

      This is a starter note. Replace it with one of your real opening patterns.
    tags:
    - pattern
    container_path: Writing Style Guide > Patterns
  - title: Review checklist
    body: |
      What to flag when reviewing an existing draft against the style guide. Replace with your real checklist.

      - Banned words or phrases present.
      - Banned punctuation present (em dashes, en dashes).
      - Banned structures present (fake contrast, three-adjective stacks).
      - Tone drift: hedging, buzzwords, over-polished AI tics.
      - Unsupported claims: numbers without a source, customer language without a citation.
      - Pattern mismatch: opens with warm-up instead of news; closes with rhetorical question.

      When you find any of the above, rewrite the offending span and explain the change in one line.

      This is a starter note. Replace it with your real review checklist.
    tags:
    - pattern
    container_path: Writing Style Guide > Patterns

Write once. You both remember.

Free to start. No credit card required.

Give your AI a memory

Works with Claude and ChatGPT today. Gemini coming soon.