AI Writing Pattern Checker
Find the words and habits that make text read as AI-generated, so you can edit them out.
Score a prompt on the parts that change the answer, and see what is missing.
Nothing is sent anywhere. The score comes from rules you can read below, not from a model, so you can argue with it — which is rather the point.
The Prompt Analyzer scores a prompt out of 100 on eight visible checks — task, output format, context, role, length, examples, constraints and audience — and lists the vague words worth replacing, so you know exactly what to add before you send it.
Paste a prompt and get a score out of 100, built from eight checks that each map to something you can see in the text: is there a task with a verb, is the output format named, is there context the model could not infer, a role, a concrete length, an example, constraints and an audience.
The checks are weighted by how much a missing part usually costs. A prompt without a clear task or output format produces a different shape of answer every run, so those two carry the most weight; a missing tone note is a small loss. Every failed check tells you what to add, and the words that promise a quality without defining it — good, engaging, detailed, professional — are listed so you can replace them with a number or an example.
Nothing here calls a model, which is deliberate. The score is reproducible, it is the same every time for the same text, and you can read exactly why it is what it is — which makes it something to argue with and improve against rather than a verdict from a black box.
Rule-based checks give the same score for the same text every time, so improvements are measurable.
A missing task or output format costs far more than a missing tone note, and the score reflects that.
Each failed check explains what to add and why it changes the answer.
Adjectives such as detailed, engaging and professional are listed so you can make them measurable.
Asking for something brief and comprehensive at once is flagged.
Any prompt for ChatGPT, Claude, Gemini or an API call. The score updates as you edit.
Each failed check explains what is missing and why it matters.
Swap adjectives like "detailed" for something measurable, such as a word count or three named items.
Task first, then format, then an example. The score rises fastest that way.
Eight rule-based checks, each with a weight: a clear task (25), a named output format (18), context (15), a role (10), a concrete length (10), an example (8), constraints (8) and an audience or tone (6). The score is the sum of the checks that pass.
No. It means the prompt has the parts that make answers predictable. Whether the task itself is well chosen, and whether the context is correct, is something only you can judge.
They mean different things to you and to the model. "At least three specific figures" is enforceable; "detailed" is a guess. Replacing them is usually the quickest improvement available.
Not measurably with current models. It is harmless, but it costs tokens and it is not what makes a prompt work. Specificity is.
No. The analysis is JavaScript running in your browser, and nothing is sent to a model or a server.
What people who use the Prompt Analyzer 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 Prompt Analyzer. It takes half a minute and no account.
Write the first reviewFind the words and habits that make text read as AI-generated, so you can edit them out.
Work out what an AI feature will cost per call, per day and per month.
Find the right resolution for SDXL, FLUX, DALL-E or Midjourney at any aspect ratio.
Fill a prompt template from spreadsheet rows and get one prompt per row.
Split long documents into overlapping chunks for embeddings and RAG.
Turn conversations.json from a ChatGPT or Claude export into readable Markdown files.
See the file size of every GGUF quant of a model, and which ones fit your memory.
Convert between tokens, words, characters, pages and reading time.