NoLeak — Privacy Sanitizer for AI

Use AI and keep your privacy. Mask sensitive client data, emails, and API keys directly inside ChatGPT and Claude.

Stop leaking confidential client data, trade secrets, and API keys into AI models. NoLeak is a zero-knowledge privacy extension that lets you safely use ChatGPT and Claude without exposing sensitive information. 🔒 100% CLIENT-SIDE — ZERO DATA TRANSMITTED Unlike other tools that route your prompts through external servers, NoLeak runs 100% locally in your browser. Your sensitive records, client names, and keys never touch our servers or third-party networks. ✨ KEY FEATURES: • Instant In-Page Scrubbing: Hit Alt+S (or click the NoLeak badge) right inside ChatGPT or Claude to instantly mask sensitive data before hitting enter. • Reversible De-Anonymization: When the AI replies, click "Restore Real Names" to seamlessly restore original names, emails, and numbers with one click. • Comprehensive Entity Detection: - API Keys & Secrets (OpenAI, AWS, GitHub, JWTs, Bearer tokens) - Credit Cards (Luhn-checksum validated) & SSNs - Personal & Corporate Email addresses - Domestic & International Phone numbers - IPv4 addresses & Physical Street addresses - Client & Patient Names - Corporate Entities (Inc, LLC, Corp) - Financial & Transaction Amounts • Session Vault: Inspect, export, or wipe your active token keymap at any time. 🛡️ DESIGNED FOR: - Legal Professionals & Paralegals protecting attorney-client privilege - Software Engineers & DevOps preventing credential leaks in logs - Healthcare workers handling non-identifiable clinical summaries - Freelancers and contractors handling confidential client documents HOW IT WORKS: 1. Type your prompt in ChatGPT as usual. 2. Press Alt+S or click the green NoLeak shield. 3. Your sensitive names and keys are replaced with safe semantic tokens (e.g. {{PERSON_1}}). 4. Submit the prompt to the AI. 5. In the AI response, click "Restore Real Names" to get your final clean text. PRIVACY GUARANTEE: NoLeak does not collect telemetry, personal identifiers, or prompt contents. Zero tracking. Zero logs. Support & Inquiries: adhamkrvmpi@gmail.com