Avoid Ai Writing
Skill that audits and rewrites content to remove AI writing patterns. Use it with your favorite agents including Claude Code, OpenClaw, and Hermes.
Last verified:
What is Avoid Ai Writing?
Avoid AI Writing is a free, open-source writing tool that audits and rewrites content to eliminate AI writing patterns ("AI-isms"). It identifies 36 categories of structural patterns that AI writing produces—such as em dash overuse, hollow intensifiers, formulaic transitions, uniform paragraph lengths, and hedging language—then maps them against a 109-word replacement table to flag exactly what sounds like AI and why. Unlike typical AI detectors that output a score, this tool produces clean text that reads like a human wrote it, with a detailed breakdown of every change and the rationale behind it.
The tool operates in two modes: Rewrite mode (default) flags AI patterns and rewrites the text, then runs a second-pass audit to catch any lingering patterns; Detect mode flags patterns without rewriting. It includes a free Chrome browser extension that runs scanning entirely locally in your browser with no text transmitted to servers, offering unlimited scans, 36 pattern categories, color-coded severity levels from P0 (critical chatbot leakage) to P3 (stylistic), and a floating scan button on every textarea. The paid deep audit feature provides full Claude-powered rewriting with second-pass cleanup, detailed change diffs, and pay-per-audit via $avoid token burn or credit pack.
Avoid AI Writing is ideal for content creators who want to audit their work for unintentional AI-like patterns, educators verifying student submissions, content managers ensuring authentic brand voice, journalists fact-checking sources, and AI agent operators using Claude Code, OpenClaw, or Hermes. The detection engine is MIT-licensed and open source, allowing users to understand exactly what patterns trigger flags. The tool integrates with agentic frameworks via a portable SKILL.md file and supports slash commands for Claude Code and clawhub installation for OpenClaw.
Avoid Ai Writing pricing
Pricing model: Freemium
Free tier includes the Chrome browser extension with client-side pattern detection across 36 categories, instant results, no account required, no text transmitted, unlimited scans, score plus flagged issues plus suggestions. Deep audit (paid) includes everything in the free extension plus full Claude-powered rewrite, second-pass cleanup on the rewrite, detailed change diff with rationale, and pay-per-audit via $avoid token burn or credit pack. Each audit burns AVOID token at a fixed USD cost of approximately $0.25 regardless of token price. The AVOID token is a Solana SPL token used to gate paid audits, with tokens burned permanently per audit.
Avoid Ai Writing pros
- Free Chrome extension runs entirely locally in browser
- No text leaves your browser unless you explicitly opt in
- Unlimited free scans with no account or sign-in required
- 36 pattern categories covering content, language, structure, communication, and meta patterns
- 109-entry word replacement table with 3 tiers reduces false positives
- Two-pass detection catches recycled transitions and lingering issues
- Color-coded severity from P0 critical to P3 stylistic
- Open-source MIT license with transparent detection methodology
- Works with Claude Code, OpenClaw, and Hermes agents
- Floating scan button appears on every textarea you focus
- Keyboard shortcut Ctrl+Shift+A scans focused field instantly
- Detect-only mode for auditing without rewriting
- Detailed change diff with rationale for every replacement
- Popup paste-and-scan for quick checks on copied text
- Per-site disable list and toggle for floating button in options
Avoid Ai Writing cons
- Requires technical comfort to install from GitHub repository
- No polished web interface for non-technical users
- Can produce false positives with formal human writing styles
- Deep audit requires payment via $avoid token or credit pack
- Extension only available for Chrome browser
- No enterprise features like API integration or team dashboards
- Mobile version still in Android beta
- Detection logic may need updates as new AI patterns emerge
Frequently asked questions about Avoid Ai Writing
What is Avoid AI Writing?
Avoid AI Writing is an open-source skill that audits and rewrites content to remove AI writing patterns. It identifies 36 categories of structural patterns that AI produces and uses a 109-word replacement table to flag what sounds like AI and why, outputting clean human-like text with detailed change breakdowns.
How does the browser extension work?
The free Chrome extension runs scanning entirely in your browser's JavaScript engine. A floating scan button appears on every textarea you focus—one click runs the scan. No text is transmitted, logged, or stored on any server unless you explicitly click the deep-audit button. It checks 36 pattern categories locally with instant results.
Is my text private when using the extension?
Yes. The extension runs scanning entirely locally in your browser. No text is transmitted, logged, or stored on any server. Your text only leaves your browser when you explicitly click the deep-audit button, and even then it's sent via chrome.storage.session plus same-origin postMessage so it never appears in the URL or browser history.
What are the 36 pattern categories?
The categories cover content patterns (significance inflation, promotional language), language patterns (copula avoidance, synonym cycling), structure patterns (em dash overuse, inline-header lists, uniform paragraph lengths), communication patterns (chatbot artifacts, sycophantic tone, cutoff disclaimers), and meta patterns (rhythm uniformity, over-polishing, hollow intensifiers).
What is the difference between Rewrite mode and Detect mode?
Rewrite mode is the default—it flags AI patterns, rewrites the text, prints a changelog, then runs a second-pass audit to catch patterns that survived the first edit. Detect mode flags AI patterns without rewriting, grouping flags by severity (P0/P1/P2) and distinguishing real problems from judgment calls. Trigger Detect with 'detect', 'flag only', or 'scan'.
How does the word replacement table work?
The 109-entry table has 3 tiers. Tier 1 words are always flagged as AI vocabulary. Tier 2 words flag when they cluster together. Tier 3 words flag only at high density. This tiered approach reduces false positives while maintaining high detection rates for AI tells like 'leverage', 'seamless', 'watershed moment', and 'serves as'.
Which AI agents are compatible?
Avoid AI Writing works with Claude Code, OpenClaw, Hermes, and any agentskills.io-compatible agent. It's packaged as a SKILL.md file with platform frontmatter. For Claude Code, run 'git clone https://github.com/conorbronsdon/avoid-ai-writing ~/.claude/skills/avoid-ai-writing'. For OpenClaw, use 'clawhub install avoid-ai-writing'.
What is $avoid token?
$avoid is a Solana SPL token used to gate the paid deep audit feature. Each audit burns AVOID at a fixed USD cost of approximately $0.25 and removes those tokens from circulation permanently. The token burns into Solana's token burn mechanic, and the detection engine ties to this via the web app at avoidaiwriting.com.
Can I modify the detection parameters?
Yes. The tool is fully open-source under the MIT license. Users can fork the GitHub repository, modify detection parameters, contribute improvements to the community, or adjust the 36 pattern categories and 109-word replacement table. The detection engine is MIT-licensed and the methodology is fully visible in the repository.
Does Avoid AI Writing work with mobile devices?
There is an Android mobile beta available at avoidaiwriting.com/mobile. The free Chrome extension works on desktop browsers, and the paid web app at avoidaiwriting.com returns deeper Claude-powered audits and rewrites. Mobile support is still in beta development.