Token Counter
Estimate how many tokens your text costs before you send it.
Find the words and habits that make text read as AI-generated, so you can edit them out.
This is not an AI detector, and nothing honest can be. It counts the habits that generated prose leans on so you can edit them out — every one of them is something a person can write too. Nothing is uploaded.
The AI Writing Pattern Checker highlights the vocabulary, connectives, hedges, openers and structural habits common in unedited AI-generated text, measures how much sentence length varies, and shows where to edit — without pretending to be an AI detector.
Readers have learned what unedited chat-model prose sounds like: it delves, it navigates landscapes, it is not just a tool but a game-changer, and it ends with in conclusion. Once someone notices those habits, they stop trusting the rest of the text — whether or not a person wrote it.
Paste any draft and the checker counts the patterns in four vocabulary groups — overused words, essay connectives such as moreover and furthermore, hedging and throat-clearing, and assistant openers left in from a chat — and six structural habits: em dashes, rule-of-three lists, not only but also, the it is not just X, it is Y antithesis, bolded lead-ins on every bullet, and emoji used as section markers. It also measures how much your sentence lengths vary, because a steady, uniform rhythm is one of the clearest signs of a draft nobody reworked.
This is deliberately not an AI detector. Detectors that output a percentage are unreliable and regularly flag human writing. Instead you get the exact words and places worth a second look, highlighted in your text, with a rate per thousand words and practical ways to fix them.
Overused words, essay connectives, hedging and chat openers, each with counts.
Em dashes, rule-of-three lists, not only but also, antithesis, bold lead-ins and emoji headers.
Flags a flat, uniform sentence length, one of the clearest signs of unedited text.
Every flagged term is marked in your own text so edits are quick.
No unreliable percentage — just specific, checkable places worth a second look.
An article, an essay, an email or a post. Longer texts give steadier rates.
Each group lists the terms found and how often, with why the pattern stands out.
Flagged vocabulary is highlighted in amber and em dashes in red, in place.
Cut openers and conclusions, use plain words, and vary your sentence lengths.
No, and it does not claim to be. It counts writing habits common in generated text so you can edit them. Every pattern here is also something people write, so a high count is a prompt to revise, not proof of anything.
Common examples are delve, tapestry, testament, realm, landscape, crucial, robust, seamless, leverage and foster, along with moreover, furthermore and phrases such as it is important to note.
Chat models use em dashes far more often than most people typing quickly. One or two are normal; several per paragraph is a recognisable habit.
The spread of sentence lengths. Human editing tends to mix short and long sentences, while unedited generated text often keeps a steady middle length.
No. The check runs in your browser and nothing you paste is stored or sent.
What people who use the AI Writing Pattern Checker think of it. Reviews come from visitors and are posted as written.
No ratings yet
Yours would be the first.
0 reviews
Newest first
No reviews yet
Be the first to rate the AI Writing Pattern Checker. It takes half a minute and no account.
Write the first reviewEstimate how many tokens your text costs before you send it.
Write valid tool-calling JSON schemas without hand-editing braces.
See the file size of every GGUF quant of a model, and which ones fit your memory.
Generate robots.txt rules to block GPTBot, ClaudeBot, CCBot and other AI crawlers.
Score a prompt on the parts that change the answer, and see what is missing.
Write both custom instruction boxes, each checked against the 1,500 character limit.
Split a dataset into training, validation and test files with a repeatable shuffle.
Turn input and output pairs into a clean, consistent few-shot prompt block.