🤖

IA et confidentialité Web ✓ Prêt à l'emploi

Anonymiseur & Restaurateur de Prompts IA

Protection bidirectionnelle pour ChatGPT et Claude : remplacez données confidentielles par des jetons et restaurez en 1 clic.

Sanitiseur IA Bidirectionnel Masquage [ ] Crochets Conforme RGPD 100% Navigateur

🤖

AI Prompt Anonymizer & Restorer

Two-Way Privacy Protection: Auto-mask confidential names, companies, and [brackets] before AI, restore with 1 click.

💡

Quick Masking Shortcut: To manually hide any confidential word or phrase, simply wrap it in brackets [ ] (e.g., [Sarah Jenkins] or [TechNova GmbH, Berlin]). The tool will automatically anonymize it and restore it after AI responds!

Original Prompt (Confidential)0 chars

Safe Anonymized Prompt (Copy to ChatGPT / Claude)

Active Memory Token Map (Safe Dictionary):No tokens generated yet. Type text or use [brackets] above.

📖How to Use This Tool (Step-by-Step Guide):

1️⃣ Step 1: Paste Your Prompt

Paste your prompt. Wrap any confidential client names or company identifiers in [ ] brackets (e.g. [TechNova GmbH]). The tool instantly tokenizes them.

2️⃣ Step 2: Prompt ChatGPT or Claude

Click "Copy Safe Prompt" and paste it into ChatGPT. The AI only sees placeholder tokens like [CUSTOM_1], keeping your confidential data 100% secure.

3️⃣ Step 3: Restore with 1 Click

Copy the AI's answer, switch to the "2. Restore AI Response" tab, paste it, and click "Restore Real Names". All original identities are instantly restored!

Privacy Tip 100% In-Browser & Offline

🤖

Mask Client PII Before Pasting into ChatGPT

Two-way client-side token anonymization and 1-click restore.

Try AI Anonymizer →

⚙️ Comment cet outil fonctionne localement

Lorsque vous utilisez cet outil sur PrivacyKit, chaque octet est traité dans la mémoire vive de votre appareil.

1. Aucun Trafic Réseau: Aucune requête HTTP POST ou PUT n'est envoyée à un serveur distant.
2. APIs Web Natives Modernes: Utilisation de HTML5 Canvas, Web Cryptography (SubtleCrypto) et parseurs binaires client.
3. Empreinte Zéro: Fermer l'onglet libère instantanément la mémoire RAM. Aucun cookie persistant.

💡 Foire aux questions (FAQ)

How does manual bracket [ ] masking work?

Simply write any word or phrase inside square brackets like [John Doe] or [Secret Design Studio]. The tool automatically converts it into a safe token like [CUSTOM_1] before sending to AI, and restores the original word when you paste the AI reply back.

How does the Two-Way AI Prompt Anonymizer work?

The tool substitutes confidential names with structured placeholders like [PERSON_1] and [COMPANY_1] while preserving grammatical sentence context. Once ChatGPT generates the response, pasting it back restores the original names instantly.