Suppresseur de schémas d'écriture IA

Remplir ce prompt

Réécrit un texte pour supprimer les schémas typiques de l'écriture générée par l'IA tout en conservant intactes toutes les affirmations du texte original.

Modèle de prompt

Act as a writing editor who removes the visible signs of AI-generated text.

Rewrite this text so it reads as if a person wrote it:

{{Text: use the attached text}}

Hard rules:
- Invent nothing. No fact, name, number, date, quote or citation that is absent from the source. If a sentence needs a detail you do not have, write the plain version without it.
- Keep every claim, but not the shape. Compress dull stretches, merge or split paragraphs, dwell where a human would.
- Use no em dashes and no en dashes. Replace each with a period, a comma, a colon, parentheses or a restructured sentence.
- Match the register of the source: formal, casual or technical. Add personality only where the content calls for it. In reference, legal or technical text, plain and neutral is the human voice.
- Rewrite prose only. Leave code blocks, frontmatter, data, quotations, titles and link targets untouched.

Remove these patterns:
- Inflated significance: stands as, serves as, is a testament to, pivotal, underscores its importance, reflects a broader shift, evolving landscape.
- Padded participles tacked onto sentences: highlighting, ensuring, reflecting, fostering, showcasing, contributing to.
- Promotional language: vibrant, rich, breathtaking, nestled, renowned, boasts, commitment to.
- Vague attribution: experts argue, observers have cited, industry reports, some critics say. Name the source or cut the claim.
- Copula avoidance: write is instead of serves as, represents, features, offers.
- Negative parallelism: not only X but Y, it is not just X it is Y, and clipped tails such as no guessing.
- Forced groups of three, synonym cycling for one subject, false ranges from X to Y on no real scale.
- AI vocabulary: delve, crucial, key, intricate, interplay, garner, tapestry, testament, underscore, enhance, align with.
- Filler and hedging: in order to, due to the fact that, at this point in time, it is important to note, could potentially possibly.
- Signposting and chat residue: let us dive in, here is what you need to know, I hope this helps, would you like me to.
- Boldface sprinkling, inline-header bullet lists, title case headings, emoji, curly quotes.
- Aphorism formulas such as X is the Y of Z, manufactured punchlines, staccato drama, generic upbeat conclusions. End on the last concrete fact.
- Uniform hyphenation in predicate position: the report is high quality, but keep a high-quality report.

Preserve what marks a human: unusual concrete detail, mixed feelings, asides and self-corrections, uneven sentence length, first-person choices the writer can defend. Never edit a watched phrase inside a quotation, a title or a proper name.

Work in a loop: draft, then ask what still reads as machine written and whether you added any fact absent from the source, then revise. Scan the result for dashes before you finish.

Output the final rewritten text, then a short bulleted list of the patterns you removed. Nothing else.

Guide du prompt

Idéal pour

  • Partir d'un prompt structuré et réutilisable plutôt que d'une conversation vierge.
  • Adapter le prompt à vos propres données tout en conservant l'intention d'origine.
  • Les workflows liés à Writing & Blogging.

Comment l'utiliser

  1. Copiez le modèle de prompt complet.
  2. Remplacez chaque variable par votre propre contexte ou choisissez l'une des options suggérées.
  3. Collez le prompt complété dans ChatGPT, Claude, Gemini ou un autre assistant IA.

Variables de ce prompt

Text

use the attached text

Exemple de départ

Act as a writing editor who removes the visible signs of AI-generated text.

Rewrite this text so it reads as if a person wrote it:

use the attached text

Hard rules:
- Invent nothing. No fact, name, number, date, quote or citation that is absent from the source. If a sentence needs a detail you do not have, write the plain version without it.
- Keep every claim, but not the shape. Compress dull stretches, merge or split paragraphs, dwell where a human would.
- Use no em dashes and no en dashes. Replace each with a period, a comma, a colon, parentheses or a restructured sentence.
- Match the register of the source: formal, casual or technical. Add personality only where the content calls for it. In reference, legal or technical text, plain and neutral is the human voice.
- Rewrite prose only. Leave code blocks, frontmatter, data, quotations, titles and link targets untouched.

Remove...

Comment fonctionnent les variables

Syntaxe des variables

Les variables sont entourées de {{ et }} et suivent ce modèle :

{{variable name: option1, option2, option3 }}

Variables prédéfinies

Une sélection peut référencer une liste de variables prédéfinie à l'aide de crochets. Elles apparaissent en [orange] et fournissent des valeurs courantes comme des couleurs, des tons ou des langues.

{{Tone: [tones] }}

Listes de choix personnalisées

Vous pouvez aussi fournir une liste de choix en ligne, séparés par des virgules.

{{Format: bullet points, paragraphs, numbered list }}
💡

Astuce : vous n'avez pas besoin de l'application PUCO pour utiliser ces prompts ! Copiez simplement le modèle et remplacez chaque section {{…}} par votre propre texte directement dans ChatGPT, Claude, Gemini ou tout autre assistant IA.

Utilisez ce prompt en quelques secondes, pas en minutes

Installez PUCO pour Mac : appuyez sur un raccourci dans n'importe quelle app, remplissez un formulaire intelligent, collez dans ChatGPT, Claude ou Gemini. Chaque variable devient un menu déroulant soigné. Vos dernières valeurs sont mémorisées.

Télécharger sur le Mac App Store