PromptHush

A local-first privacy firewall for AI prompts.

PromptHush — Privacy firewall for AI prompts Every time you use a browser-based AI assistant, you risk sending API keys, client names, tax IDs, medical records, and passwords to a third-party server. PromptHush stops that before it happens. The extension scans your prompt locally in your browser the instant before you hit send. If it finds sensitive data, it flags it and replaces it with safe placeholders. The AI gets the context it needs. Your private details never leave your device. WHAT IT CATCHES - API keys, tokens, and secrets - Passwords and credentials - Social Security Numbers - Tax IDs and EINs - Credit card numbers - Medical information (diagnoses, record numbers, dosages) - Legal and case details (case numbers, privilege markers, settlements) - Email addresses - Phone numbers - Physical addresses - Confidential markers ("internal only", "under NDA", "trade secret") - Client and company names - Person names - Financial figures and compensation data HOW YOU USE IT 1. Type your prompt normally in any supported AI tool. 2. PromptHush scans the text before it sends. 3. If sensitive data is found, you choose: auto-redact, replace with placeholders, or override. 4. The AI receives a clean prompt. Your private details stay on your machine. WORKS WHERE YOU WORK Runs automatically on popular AI assistants you already use in your browser. No setup needed — install and it starts protecting immediately. PRIVACY FIRST All scanning happens locally in your browser. Your prompt text is never sent to our servers. We only track usage counts for your plan — never the content of what you type. WHO IT'S FOR Anyone who handles confidential information and uses AI tools daily — legal, development, recruiting, consulting, healthcare, and finance professionals. PLANS AND PRICING Free tier available with 30 scans per month. Paid plans unlock custom terms, detection presets, exportable audit logs, and more. Visit https://prompthush.com for full plan details and pricing.