SafeInput AI – Real-Time Data Leak Prevention

Protect confidential information, credentials, and sensitive data from leaking into prompts. 100% local in browser.

SafeInput AI protects your confidential data, credentials, and personal information from accidental leakage when drafting prompts in your browser. When you paste code, configuration files, or notes into web-based AI tools, sensitive details like API tokens or personal data can be exposed. SafeInput AI inspects prompts locally and provides instant masking options before the prompt is submitted. FEATURES • Real-Time Interception — Automatically pauses submission when sensitive data patterns are detected. • Cloud & API Credentials — Recognizes common API key structures, access tokens, and private keys. • Personal & Financial Data — Detects bank account numbers (IBAN), credit card formats, and identification numbers with checksum verification. • Smart Redaction — Choose between structured labels (such as [FILTERED_API_KEY]), character masking (XXXX), or confidential markers. • Custom Keywords — Define your own confidential project codenames, client names, or internal terms to block. • Interactive Test Sandbox — Test prompts directly in the popup interface without sending anything. HOW IT WORKS 1. Write or paste your prompt normally in any supported web interface. 2. If sensitive data is detected upon hitting Send or Enter, a secure dialog appears with detected items. 3. Choose to send a sanitized version, insert masked text for manual editing, or proceed unshielded. PRIVACY FIRST • 100% Local Execution — All pattern checks, algorithms, and regex operations run entirely on your device. • Zero Data Collection — No prompt data, personal info, or logs are ever transmitted to external servers. • No Tracking — No analytics, third-party cookies, or remote telemetries. SUPPORT & FEEDBACK Questions, feature requests, or issue reports can be submitted via our official GitHub repository: https://github.com/rubenwinkelman/safeinput-ai