Media
5 items
Description
Scan prompts locally, redact secrets and personal data, guard clipboard pastes, and keep a private audit trail. Stop secrets and personal data before they slip into an AI chat. Zovo Redact checks the active prompt composer locally and gives you a clear choice before a risky send: redact the matching values, send once, or cancel. The extension never uploads prompt text for detection. WHAT IT PROTECTS • Common API key and access token shapes • Private key blocks and credential assignments • Email addresses, phone numbers, IP addresses, and payment-card candidates • Large clipboard pastes and credential-shaped clipboard text HOW IT WORKS • Scan the current draft at any time • Use one-click “Redact & send” when the pre-send guard finds sensitive text • Set a separate fresh-chat timer for Gemini, ChatGPT, Claude, and Grok • Review a local audit trail showing category names and provider UI • Export the audit trail as JSON PRIVATE BY DESIGN Detection happens in your browser. Audit entries contain only category metadata, provider, time, and action. They never contain prompt text or the values that matched. A timer opens a fresh chat in the current tab; provider-side retention policies still apply. WORKS WITH Gemini, Google AI Studio, ChatGPT, Claude, and Grok. The optional Pro path adds longer timers and complete audit export. The core pre-send scan, redaction, and paste warning remain available without Pro. This extension is open source under GPL-3.0. Complete modified source: https://github.com/theluckystrike/zovo-redact Privacy policy: https://zovo.one/privacy-policy Pricing: https://zovo.one/pricing