Paste a prompt and see how complete it is. The tool checks for the six parts that make a prompt reliable (role, task, context, output format, constraints and examples), warns about vague words, estimates the token count, and rewrites the layout with your own sentences sorted under labelled headings. It works from a fixed checklist of rules, not a language model, so the same prompt always gets the same result and nothing you write is sent anywhere.
A role tells the reader whose point of view to take: a senior copywriter, a patient tutor, a careful editor. A task states the single thing to do, in a sentence that starts with a verb. Context supplies what is not obvious, such as the product, the audience and the reason. Together these three decide most of the quality.
The other three make the answer usable. An output format says whether you want bullets, a table, JSON or two paragraphs, and how long. Constraints say what to avoid and what limits apply. An example shows rather than tells. A prompt that has all six leaves little to chance, while a one-line request leaves the reader to guess at every one.
Words like good, nice, better, some and interesting feel specific when you write them, because you know what you mean. The reader does not. 'Make it better' could mean shorter, warmer, more formal or more detailed, and each guess is a coin toss. Every vague word is a place where the result can drift from what you pictured.
The fix is almost always a number, a name or a sample. Instead of 'a few ideas', ask for 5. Instead of 'a nice tone', give two adjectives and one thing to avoid. Instead of 'some context', paste the two facts that matter. Prompts get shorter to read and more exact when each fuzzy word is traded for something checkable.
No. It is a rule-based checker. It looks for keywords and patterns that signal each part of a good prompt, counts words, and sorts your sentences under headings. It never rewrites your wording or adds content of its own, apart from empty bracketed slots for the parts you left out.
Each part found adds points: task 25, context 20, output format 20, constraints 15, role 10, examples 10, for 100 in total. Every vague word takes off 2 points, up to 10, and a prompt under 10 words is capped at 40 because it cannot carry much detail. For example, a prompt with a task and context only scores 45, and two vague words bring it to 41.
As the number of characters divided by 4, rounded up. That is a common rule of thumb for English text. Real counts depend on the model and the language, so use it to judge size, not to bill exactly.
Detection is keyword based. If you cover a part in unusual wording, the rule may not see it. Add a plain cue such as 'Format:', 'Audience:' or 'Do not' and it will be recognized. The checklist is a prompt for you to look, not a verdict.
No. The checking happens in your browser and your text is not uploaded.
Related: Character counter, Readability checker, Headline analyzer.
Start for free or see all free tools. We.Inc is free to start, with no card needed. AI building is on paid plans from $20 a month.
We.Inc is an AI-powered website builder you can resell under your own brand. Launch a branded client dashboard, bill on Stripe Connect, and deliver AI-generated websites in minutes. White-label plans start at $99 a month for 25 client sites, with a 7-day free trial and no per-site fees.