Dev.to
7/6/2026

The original title is "I Built a Privacy Scrubber So My Prompts Stop Leaking Data to AI Models"
Original: I Built a Privacy Scrubber So My Prompts Stop Leaking Data to AI Models
Short summary
Addresses critical data-leakage risk: developer built a client-side tool that automatically scrubs PII, corporate data, and regulated content from prompts before they're sent to AI models. Uses regex-based detection with encrypted local vaults so real values can be restored from AI responses. Entirely browser-based, open-source, with roadmap for NER detection and extension.
- •Client-side privacy scrubber removes PII and corporate data from prompts before AI model submission
- •Regex-based detection with encrypted local mapping of real values for response restoration
- •Open-source, browser-only operation, roadmap includes NER and browser extension
Generated with AI, which can make mistakes.
Is this a good recommendation for you?



