SafePrompt

Redacts secrets and PII in AI prompts. 100% local, free.

SafePrompt protects your AI prompts by redacting secrets before you send them. When you type or paste into supported AI chat sites, SafePrompt scans the composer for sensitive patterns — API keys, tokens, emails, phone numbers, credit cards, and private keys — then replaces them with placeholders like [API_KEY], [PII], or [SECRET], or blocks send in block mode. Everything runs locally in your browser. SafePrompt does not upload your prompts or secrets to our servers. No account required. Features • Local detection and redaction — no cloud processing • Undo / Allow once toast after each redaction • Balanced and Strict presets • Redact mode or block mode • Local stats for detections, redactions, and categories (no secret values stored) • Whitelist exact phrases you choose to allow • Works on supported AI chat websites listed in the extension permissions What SafePrompt can detect • Common API keys and access tokens • PII (emails, phone numbers, valid credit card numbers) • Secrets such as private keys and contextual credential strings Important SafePrompt reduces accidental leaks. It cannot catch every secret. Always review prompts before sending sensitive data. Support: safeprompt.help@gmail.com Website: https://safe-promt.web.app Privacy: https://safe-promt.web.app/privacy Terms: https://safe-promt.web.app/terms