Why a borrowed baseline gives you the wrong answer
Telling you how AI-flavoured your writing reads means comparing it to real writing in the same language. Comparing it to the wrong language does not give you a smaller signal, it gives you a meaningless one dressed up as a real measurement.
NeverPrompted measured its English baseline from contemporary English prose, and every document in that corpus was verified to be English before it was used.
The rewrite itself works the same way: it targets the phrasing habits that read as AI-generated in that specific language, not a translated list of English ones.
If the language cannot be worked out
When the engine cannot confidently tell what language a document is in, it says so rather than guessing and quietly measuring against the nearest match. A wrong-language result looks exactly like a right one until you notice everything you write scores the same, which is the sign something upstream is broken.
You can also set the language explicitly before running the check, if you know it and the engine is unsure.
What the result can and cannot tell you
A result for English tells you how this draft compares with the English reference used by the engine. It does not grade your fluency, decide whether the writing is good, or identify who wrote it. A formal register, specialist subject or careful second-language edit can all sit away from the reference for perfectly ordinary reasons.
Use the highlighted passages as editing prompts, not verdicts. Keep a change when it makes the sentence clearer or more like something you would say; ignore it when it changes your meaning, your terminology or a deliberate choice of tone. The point is control over the draft, not a number to chase.
The rewrite is equally explicit about its limits. It can change stock phrasing and rhythm in English while preserving the facts you supplied, but it cannot prove authorship or guarantee a result against a vendor's private detector. You see the before-and-after text and decide what stays.
That distinction matters when a draft is formal or translated. A lower distance from the reference is not automatically better, and a higher distance is not a failure. The useful question is whether a suggested change improves this sentence for its intended reader.
Keep the draft you started with, review each proposed change, and treat the measurement as one piece of editing evidence. Your judgement about meaning, tone and context remains the final check.
Wherever this page describes a result: a detected mark is not proof of authorship, and an absent mark is not proof of human authorship. NeverPrompted's on-device rewrite can reduce detectable evidence but cannot guarantee defeating a vendor's undisclosed watermark, on any tier.