Prompt to Remove Translationese From AI Output
Splits long sentences, turns passive into active, clears pronoun clutter and rebuilds the rhythm. For marketing and SEO copy.
The text below is machine-translated and reads like English sentence structure wearing English words. Turn it into something that reads like it was written in English.
TEXT: [TEXT]
TEXT TYPE: [TYPE]
TARGET READER: [READER]
MUST KEEP: [KEEP]
STEP 1 — MARK THE TELLS
Find and quote examples of:
- Long sentences with three or more subordinate clauses
- Passive constructions stacked on top of each other ("is being implemented", "has been provided")
- Translated connectives ("In this respect", "With regard to the aforementioned", "It should be noted that")
- Noun-stack translations ("user experience optimisation" instead of "improving the user experience")
- Redundant pronouns ("it", "this", "these" where English would drop them)
- Verb inflation: "carry out", "perform", "implement", "facilitate" where one verb would do
STEP 2 — REWRITE
- Passive to active; name the actor. If unknown, one passive per paragraph maximum.
- Split long sentences. In English, average 15-20 words.
- Turn noun stacks back into verbs: "optimisation" → "improve".
- Drop "this/these" and repeat the noun instead.
- Keep everything in MUST KEEP exactly (brand, technical terms, legal wording).
STEP 3 — COMPARE
Give the five most important changes side by side: before / after / why.
Score: rate the new text out of 10 for natural English, and say in one sentence why it is not a 10.
Best result: include the source-language original as well. The model then rebuilds meaning from the source instead of inheriting the translated structure.
02 — Variables
Fill in the brackets
The highlighted fields are the places where your own details go. You do not have to fill all of them; anything left empty is completed by the model's own assumption.
[TEXT]
[TYPE]
[READER]
[KEEP]
03 — How to use
How to use it, step by step
- 01
If you can, give the original text too and say "work from the original, not the translation". That is where the biggest quality gain is.
- 02
Do not skip step 1 — naming the tells is the fastest way to see your own repeated patterns.
- 03
Do not be surprised if it scores a 10; models are generous with their own output. A 7 or 8 with a reason is more useful.
- 04
Read the rewrite aloud. If you run out of breath, the sentence is still too long.
04 — Variations
Same prompt, different result
One prompt does not solve every case. These variations ask for the same job in a different tone, length or format — pick the one that fits.
From the original
Translate the English text below into [TARGET LANGUAGE]. Rules: (1) work paragraph by paragraph, not sentence by sentence — meaning first; (2) you may reorder sentences for natural flow; (3) do not translate idioms literally, use an equivalent; (4) keep technical terms. Text: [TEXT]
The highest-quality route.
Industry register
Rewrite the same text in [INDUSTRY] terminology. Keep terms that are standard in the field, simplify invented terms that the reader would not recognise. For every term you keep, give a one-line justification. Text: [TEXT]
Legal, medical and technical text.
05 — Tips
Notes that take the result up a level
- Stacked passives are the most visible sign of translated text. Replace each one with an active verb.
- "This allows", "this enables", "furthermore" can all usually just be deleted — the text gets shorter and stronger.
- A translated text often has near-identical sentence lengths. Break that pattern deliberately.
06 — FAQ
Frequently asked about AI Auditing
Isn't a good translation engine enough?
Modern engines are fluent but preserve sentence boundaries: a long source sentence stays long in the output. Natural English often needs that sentence split, and the target language's own rhythm. That is a rewriting job, not a translation job.
Why is translated text so recognisable?
Because the underlying grammar differs. Translating word by word fills one language's skeleton with another's vocabulary. Readers feel something is foreign without being able to name it. This prompt changes the skeleton.
Does translationese hurt SEO?
Indirectly, yes. Dwell time and engagement drop because readers leave in the first paragraph, and search engines read that as a quality signal. It also increases duplicate-content similarity with the source version.
Next
You may also find these useful
Prompt to Catch Fake Citations: When the AI Invents Sources
Tests every source on four axes (does it exist, does it support the claim, is the metadata consistent, is it accessible) and surfaces the suspect ones first.
Intermediate ChatGPTClaudeAI Plagiarism Check: Rank Your Text by Risk Before Submission
Scans definition sentences, opening paragraphs and conclusions separately, then gives a source type and a rewrite direction for each risky sentence.
Advanced ChatGPTClaudePrompt to Make AI Text Sound Human (Remove the Tells)
Finds the clearest AI tells — rule-of-three balance, empty adjectives, cliché openers, connective overload — and rewrites around them without losing meaning.
Intermediate ChatGPTClaudeAI Misreading Your Table? Prompt for Source-Traced Numbers
Requires row and column coordinates, unit and period for every number, and marks anything unreadable as "uncertain". Essential for financial analysis.
AdvancedThis page was prepared by the CopyPromt editorial team, first published 25 October 2026. Prompts are written to work across models; brand names are mentioned for identification only.