Author: Koutian Wu
Github, LinkedIn: ktwu01
Release date: Apr 20, 2025
The AI Word and Phrase Highlighter is designed to detect common words and phrases frequently used in AI-generated content.
为啥要做：，优势
免费开源。
本地部署，数据不上传，隐私保护。
可以根据需要自定义AI词库。
某种意义上也可以作为违禁词词库查询。。
捐赠选项，捐赠支持作者。
The tools work by scanning text for words and phrases commonly used by AI models like ChatGPT, Claude, and others. These words and phrases are ranked based on their frequency in AI-generated documents compared to human-written documents. The data is drawn from research by various detection tools like ZeroGPT and GPTZero.
When analyzing text, the tools identify common AI markers like:
- Overused transition phrases ("Furthermore", "On the other hand")
- Formal academic constructions ("It's important to note", "In conclusion")
- Verbose expressions ("When it comes to", "In the realm of")
- Distinctive vocabulary ("delve", "plethora", "harness", "tapestry")
You can use these tools in several ways:
- For Content Creation: Identify and replace AI-like patterns in your writing
- For SEO: Optimize content to reduce AI detection probability
- For Education: Compare AI vs. human writing patterns
- For Research: Analyze large text corpora for AI indicators
The HTML reports provide color-coded highlighting of AI indicators and suggest alternative words and phrases to make text appear more human-written.
To use these tools:
- Git Clone
git clone https://github.com/ktwu01/AI-Text-Detector.git - Install required dependencies:
pip install pandas matplotlib seaborn streamlit sqlite3 - Run the tool
streamlit run KW-ai_word_highlighter.py
Example result: