ZeroReveal - Zero Trust PII Redaction for AI Chats

# ZeroReveal ZeroReveal is a Zero Trust privacy layer for AI chat apps. It acts like a security firewall between your keyboard and the chatbot, helping prevent sensitive data from leaving your browser. Accidentally pasting customer data, emails, phone numbers, credentials, or tokens into AI tools is easy. ZeroReveal scans your prompt locally and replaces sensitive values with placeholders like `[EMAIL_REDACTED]` and `[PHONE_REDACTED]` before sending. Works with: - ChatGPT - Claude - Gemini ## Key Features - One-Click Redaction: Instantly scans and replaces sensitive data in your prompt. - Advanced Name Detection: Smart NLP plus phrase detection to catch names like "John Smith." - Secrets Protection: Detects common API keys, tokens, JWTs, and private keys. - Smart Key/Value Redaction: Hides values like `password=...`, `token: ...`, `api_key=...`. - Connection String Safety: Redacts credentials inside database URLs. - Preview Before Replace: Review redactions before applying. - Send Without Redaction: Optional bypass when you want to proceed unchanged. - Custom Rules: Add your own regex patterns for internal IDs and formats. - Whitelist (Safe List): Exclude trusted terms from redaction. - Auto-Redact on Send: Optional protection right before sending. - Auto-Redact on Paste: Optional protection right after paste. - Auto-Save Settings: Toggles and inputs save as you change them. - Formatting Preserved: Keeps paragraph breaks and spacing intact. - Fully Local Processing: No servers, no cloud processing. ## Privacy First ZeroReveal runs locally in your browser. We do not track keystrokes, store chats, or collect analytics. Your data stays on your machine. ## Permissions - Storage: Saves your settings, whitelist, custom rules, and local usage stats. ## Disclaimer ZeroReveal is an independent extension and is not affiliated with or endorsed by OpenAI, Anthropic or Google. ## Release Notes ### v3.1 - Added compact in-chat controls with a small grouped redact button and pause/resume icon. - Improved the on-page UI so ZeroReveal takes less space beside AI chat send buttons. - Added a hidden-by-default redaction badge that appears only after redactions. ### v3.0 - Added support for Claude and Gemini. - Improved cross-platform prompt detection and send-flow reliability. - Improved release packaging and versioned distribution workflow. ### v2.1 - Added "Send Without Redaction" option in preview. - Added auto-save for settings toggles and inputs. - Improved name detection to reduce false positives (for example, "i am taking"). - Fixed formatting loss when redacting multi-line text in ChatGPT. - Improved phone matching to avoid partial-parentheses artifacts. ### v2.0 - Added secrets detection (API keys, tokens, JWTs, private keys). - Added key/value redaction (passwords, tokens, client secrets). - Added connection string redaction. - Added IP, SSN, and credit-card detection. - Added preview modal and auto-redact on send/paste. - Added custom rules and expanded whitelist support. - Added stats and session pause toggle. - Improved name detection and international phone support. ### v1.0 - Initial launch with Email, Phone, and Name detection support.