Free tools Windows power users keep installed
One-click scans. No signup required.
iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
A 40-line Python detector can flag sentence-length variation and repeated three-word phrases, but its score cannot prove whether a person or an AI wrote a passage. Its author describes using it to triage drafts and reports cases where it misled him; that is useful context, not a controlled accuracy test. Treat the script as a way to spot patterns worth reviewing—not as an authorship verdict.
What the detector measures
The example script combines two hand-selected text features into a heuristic score. It does not identify a writer, inspect a document’s history, or establish where the text came from.
Sentence-length variation (“burstiness”)
The script splits text at a period, exclamation mark, or question mark followed by whitespace, then counts whitespace-separated words in each resulting sentence. It calculates the population coefficient of variation: the standard deviation of sentence word counts divided by their mean. The implementation skips this feature when it finds fewer than three sentences.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →Here, “burstiness” means only that particular coefficient of variation. It is not a universal measure with one standard definition, and it should not be confused with perplexity, which is a different kind of language-model-based measure.
#1 Best Overall
- 🎙️ Hands-Free Voice Typing for Windows & Mac – Powered by iOS & Android dictation technology, AI VoiceWriter allows fast, accurate speech-to-text directly on your desktop. Simply speak, and your words appear in real time. Compatible with Windows 10 & above, macOS 13 & above.
- ✍️ AI Writing Assistant for Effortless Editing – Boost productivity with AI proofreading, rephrasing, and formatting. Perfect for emails, reports, creative writing, and professional content.
- 💻 Works Seamlessly in Any Desktop App – Type with your voice in Microsoft Word, Google Docs, PowerPoint, Teams, emails, and more. Just place your cursor in any text field and start speaking!
- 📱 Mobile App for Enhanced Voice Input – The AI VoiceWriter mobile app enhances voice recognition by using your phone’s microphone as an input device for clearer, more accurate dictation—while typing on your desktop. Supports iOS 15 & above, Android 9.0 & above.
- 🌎 Multilingual Voice Typing & AI Assistance – Supports 33 languages for dictation, plus AI-powered features in Chinese, English, Japanese, Korean, French, German, Spanish, Italian and, Swedish.
Repeated three-word sequences
For the second feature, the code lowercases the text, extracts runs of ASCII letters and apostrophes, and counts three-word sequences, or trigrams. It records trigram types that appear at least twice. This is a repetition signal, not a test of whether a phrase is characteristic of an AI system.
How the signals become a score
The final score gives 60% weight to capped inverse burstiness and 40% to a capped count of repeated trigram types. Those fixed weights and caps are choices made by the script’s author; they are not evidence that the score is calibrated or that a particular value corresponds to a reliable probability of AI authorship. Broader descriptions of detectors that examine text properties do not establish that commercial systems use these exact calculations.
Rank #2
- Upgraded AI-Powered Detection: Military-grade technology detects hidden cameras, listening devices, and GPS trackers with precision. Enjoy peace of mind in hotels, offices, and even your own home. Stay one step ahead of hidden threats!
- Simple, Fast & Effective: Just turn it on, sweep the area, and let the audible alarm + LED alerts notify you of threats. No technical skills needed - Press, Search, Relax! Skip expensive private investigators - protect yourself in seconds.
- Compact & Travel-Ready: Lightweight, rechargeable, and pocket-sized for discreet, on-the-go security. Toss it in your bag, purse, or pocket - perfect for travel, work, and public spaces.
- Total Privacy Protection: Don’t gamble with your security. Safeguard against spying in hotel rooms, changing rooms, offices, cars, dorms, and more. Know for sure if you’re being watched, recorded, or tracked.
- Trusted by Experts & Customers: Designed with cybersecurity and counter-surveillance professionals. Join 300,000+ satisfied users who rely on our detectors for ultimate privacy & safety.
He Harvey He presents the script as a lightweight way to sort drafts. His September 30, 2026, article is the primary source for the implementation and his anecdotes: Burstiness and n-grams: a 40-line Python AI-text detector, and where it breaks.
Where the approach breaks down
Short passages
Below three sentences, the script does not calculate its sentence-variation feature. He also reports that estimates based on only three or four sentences can swing substantially and says he treated passages under about 40 words as unknown. That roughly 40-word threshold is his practical rule of thumb, not a validated cutoff or a general standard. A score from a short sample should not be read as strong evidence in either direction.
Rank #3
- ALL-IN-ONE SCAM DETECTION – Texts, emails, videos, and QR codes all get checked automatically. Sorting real from fake stops being your job.
- KEEP SCAMMERS OUT OF YOUR WALLET – Every click is no longer a gamble. Our scam detection spots suspicious texts, email scams, SMS phishing, and fake alerts before you click.
- QR CODE SCANNING – Point the app at any code and see where it actually leads before you scan it.
- DEEPFAKE DETECTION – When a video sounds like someone you know but isn't, you hear it from us first.
- ON-DEMAND CHECKS – Got a message you're unsure about? Run it through the app and know in seconds, wherever it came from.
Formulaic writing
Legal summaries, release notes, and API documentation often follow conventions that produce repeated structures or similar sentence lengths. He reports that the script flagged such human-written material. A feature that marks regularity can therefore mistake the requirements of a genre for evidence about authorship.
Small edits can change the signal
He describes examples in which changing sentence boundaries or order, or replacing a word, removed a repeated-trigram signal without meaningfully changing the prose quality or its author. These examples illustrate sensitivity to editing; they are not a comprehensive robustness audit. They do show why a changed score should not be treated as proof that authorship changed.
Scores do not measure writing quality
A low or high score describes how the chosen measurements behave on a passage. It does not tell you whether the writing is accurate, clear, original, or suitable for its audience. Nor does this script’s score, on its own, establish that a passage is human-written or AI-generated.
What broader detector evaluations show
Independent studies add an important caution: results depend on what is being evaluated. They do not test this 40-line script, so their findings cannot be used to assign it an accuracy rate.
Best Value
- ⚛️ ACCURATE: Super sensitive semiconductor sensor. 20-year sensor life.
- 💪 SHOCK: ABS housing can take a 10ft drop.
- 🌈 DISPLAY: Large color LCD screen with temperature.
- 🚀 USES: For homeowners and tenants.
- 🕵️ TRUST: ** 1 Year Warranty ** USA Customer support in Los Angeles, California.
- BUST benchmark: A 2024 Association for Computational Linguistics benchmark assembled 25,000 texts from human writers and seven LLMs, spanning ten tasks and three sources. It evaluated five detectors and found that performance varied substantially across tasks. The study’s coverage underscores why a result on one kind of text cannot automatically be carried over to another. Read the BUST paper.
- NIST text-to-text pilot: NIST’s 2024 NIST GenAI (Pilot Study): Text-to-Text Evaluation Overview and Results, designated NIST AI 700-1 and published June 25, 2025, reports substantial variation by system and generator. Its abstract says performance varied significantly depending on the systems used; some generators could deceive most discriminators, while some discriminators could detect AI-generated content from almost all generators. This describes the pilot, not the performance of the script discussed here. Read the NIST report.
- Related metric limitations: A 2023 EMNLP paper discusses why perplexity and burstiness are often unreliable indicators of human-written text, particularly for academic writing and low-resource languages. It also addresses reference-model and out-of-domain limitations. These points support caution about predictability-based inference, but they do not directly measure this script’s error rate. Read the EMNLP paper.
The practical lesson is to evaluate a detector against the intended use: the relevant tasks and genres, the kinds of generators that matter, realistic sample lengths, and both false positives and false negatives. A benchmark that does not resemble the text population in which you plan to use a tool may not tell you how it will behave there.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How to use the score responsibly
- Use it to identify passages for review. Repeated phrasing or unusually uniform sentence lengths may be worth checking, but neither feature identifies a writer.
- Review the text itself. Check factual accuracy, clarity, appropriate sourcing, and whether the prose fits the intended audience. These are editorial checks, not conclusions supplied by the detector.
- Account for the sample and genre. A short passage gives limited sentence-level information, while technical and formal genres may be patterned by design.
- Do not make a consequential authorship decision from this score alone. The script’s author reports no controlled validation, and the broader evaluations above show that detector behavior varies with the evaluation conditions.
What the evidence does—and does not—establish
The code’s two signals and fixed weighting are clear enough to explain and inspect. He’s reported failure cases make plausible, concrete cautions about short samples, formulaic prose, and sensitivity to small edits. But the article does not report a false-positive rate or an accuracy figure for this script, and none of the cited external evaluations tests it. No numerical claim about its reliability is therefore established here.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
