How to Spot AI-Written Text at a Glance? The Truth Revealed by the 'LLM Cliché Highlighter'

A digital interface showing AI-written sentences highlighted on screen
AI Summary

Developed by Simon Willison, the 'LLM Cliché Highlighter' is a browser-based tool that detects and highlights repetitive, formulaic expressions frequently appearing in AI-written content in real time.

Imagine this: You wake up in the morning and read a newsletter on your smartphone, only to feel that the flow of the text is somehow robotic and every sentence follows the same cookie-cutter pattern. It’s that kind of writing that makes you suspect, “Did an AI write this?” Recently, we have been encountering a vast amount of AI-generated content. However, they say AI-written text has a distinct “tell.” The “LLM Cliché Highlighter,” developed by Simon Willison, is a tool designed to find exactly those traces. Source: Simon Willison Releases LLMClichéHighlighter to Detect Robotic Writing Pattern

Why does this matter?

We now consume information produced by AI every single day. However, AI’s signature bland tone and repetitive phrases can diminish the authenticity of the text and hinder readability. This tool gives writers, editors, and anyone who enjoys writing a chance to check whether their own prose is trapped in “AI-like” grammar. For those who want to fully convey their own thoughts, this tool acts as a small “filter” that helps strip away mechanical habits. Source: LLMClichéHighlighter AI Writing Cliché Detector Challenges the AI Writing Cliché Detector

Simple Explanation: An ‘AI Writing Forbidden Word’ Detector

In simple terms, you can think of this tool as an “AI writing forbidden word detector.” To use an analogy, just as we apply filters in photo apps to retouch images, this tool serves to uncover the “AI filter” layered over text. Source: LLMClichéHighlighter AI Writing Cliché Detector Challenges the AI Writing Cliché Detector

Using it is very simple. You just need to copy and paste the text you want to analyze into the website, or enter a page URL. The tool then scans the text in real time, identifies 10 common cliché patterns frequently found in AI-generated text, and highlights those sentences to make them stand out. Source: LLMclichéhighlighter, [Source: LLMClicheHighlighter Tool by Simon The AI Profit Wire](https://metadatamarketer.com/llm-cliche-highlighter-tool-by-simon-willison/)

It’s similar to a teacher using a red pen to point out clichés after you finish a writing assignment. It includes features to toggle specific patterns on and off, and even allows you to navigate between flagged sentences and view the total count, assisting with thorough proofreading. Source: tools/llm-cliche-highlighter.docs.md at main · simonw/tools · GitHub

Lightweight and Usable Anywhere

Another advantage of this tool is its accessibility. It runs directly in the web browser without requiring the installation of any heavy, bulky software. Because it operates through pattern matching within the browser regardless of your computer environment, it is light and fast enough to be used in offline environments without an internet connection. Source: LLMClichéHighlighter: детектор штампов ИИ-текстов, Source: tools/llm-cliche-highlighter.docs.md at main · simonw/tools · GitHub

Simon Willison creating this tool is a very interesting case of “self-reflection.” It is because a tool that utilizes AI has the goal of correcting the artificial linguistic habits of the AI model itself. Source: LLMClichéHighlighter: детектор штампов ИИ-текстов

How will our writing change?

The LLM Cliché Highlighter is not so much a groundbreaking technology as it is an auxiliary tool that will help mature the way we interact with AI. As technology advances, the effort to examine the content we generate and add a “human touch” will continue. This small tool, which erases the traces of AI writing, will ultimately become an important indicator that makes us rethink what writing with unique human character really is. Source: LLMclichéhighlighter by Simon Willison, [Source: Tool: LLMclichéhighlighter Simon Willison’s Weblog](https://simonwillison.net/2026/Jul/17/llm-cliche-highlighter/)

MindTickleBytes AI Reporter’s View

The situation where a tool made by AI is catching AI’s bad habits is quite paradoxical yet amusing. If this “reflective” attitude, where tools provided by technology heal the flaws of that very technology, is maintained, couldn’t AI and humans coexist in a much healthier way?

References

  1. LLMclichéhighlighter: https://tools.simonwillison.net/llm-cliche-highlighter
  2. LLMClichéHighlighter AI Writing Cliché Detector Challenges the AI Writing Cliché Detector: https://www.remio.ai/post/llm-cliché-highlighter-ai-writing-cliché-detector-challenges-the-ai-detection-industry
  3. Simon Willison Releases LLMClichéHighlighter to Detect Robotic Writing Pattern: https://aitodaybrief.com/en/news/vibe-coding/simon-willison-releases-llm-cliche-highlighter-to-detect-robotic-writing-pattern
  4. LLMclichehighlighter by Simon Willison: https://aiengineerguide.com/til/llm-cliche-highlighter/
  5. LLMClichéHighlighter Modern Orange: https://modernorange.io/item/49476802
  6. tools/llm-cliche-highlighter.docs.md at main · simonw/tools · GitHub: https://github.com/simonw/tools/blob/main/llm-cliche-highlighter.docs.md
  7. Tool: LLMclichéhighlighter Simon Willison’s Weblog: https://simonwillison.net/2026/Jul/17/llm-cliche-highlighter/
  8. LLMClichéHighlighter: детектор штампов ИИ-текстов: https://ai-uchi.ru/translations/llm-cliche-highlighter-detektor-shtampov/
  9. LLMClichéHighlighter: найти ИИ-клише в тексте ContentRun Дзен: https://dzen.ru/a/amOiPdVlSA96Ckdk
  10. LLMClicheHighlighter Tool by Simon The AI Profit Wire: https://metadatamarketer.com/llm-cliche-highlighter-tool-by-simon-willison/
AD
Test Your Understanding
Q1. How does the LLM Cliché Highlighter work?
  • Sends text to a server for analysis
  • Analyzes instantly in the web browser using pattern matching
  • Runs after installing separate software
This tool uses a pattern-matching method that runs inside the browser and can be used even offline.
Q2. What does this tool primarily detect?
  • Grammar errors
  • Formulaic, repetitive expressions often used by AI
  • Logical inconsistencies in the text
It highlights 10 cliché phrases and repetitive expressions commonly found in text generated by AI models.
Q3. Who developed the LLM Cliché Highlighter?
  • OpenAI research team
  • Simon Willison
  • Google DeepMind
This tool is a personal project developed by Simon Willison.
How to Spot AI-Written Text...
0:00