Paste any text and analyze whether it was likely written by AI (ChatGPT, Claude, Gemini) or a human. Uses heuristic analysis including sentence variation, vocabulary diversity, filler phrase detection, and burstiness scoring. 100% client-side, no data sent anywhere.
This tool analyzes multiple linguistic patterns to estimate whether text was generated by an AI model. It does not use any AI or machine learning itself -- it relies entirely on statistical heuristics that exploit known differences between human and AI writing styles.
Sentence Length Variation: Human writers naturally vary their sentence lengths -- mixing short punchy sentences with longer complex ones. AI models tend to produce more uniform sentence lengths.
Vocabulary Diversity (Type-Token Ratio): Humans use a wider variety of words relative to text length. AI often relies on a narrower vocabulary with more repetition of common terms.
Burstiness: Human writing has "bursts" of complexity followed by simpler passages. AI text tends to maintain a steady, predictable level of complexity throughout.
Filler Phrase Detection: AI models frequently use phrases like "It's important to note," "In today's digital landscape," and "Let's dive in." The density of these phrases is a strong signal.
Transition Word Density: AI overuses transition words like "Furthermore," "Moreover," and "Additionally" to create an appearance of logical flow.
Average Word Length & Syllable Patterns: AI models tend toward medium-length words and avoid both very short colloquial words and very long specialized vocabulary.
Paragraph Uniformity: AI-generated text tends to have paragraphs of similar length, while human writing naturally varies paragraph sizes.
Let us build it for you. Custom APIs, dashboards, automations — whatever you need.
Start a Project →269 free tools. No signup. Open source. Built by Christian Bucher.