PII Detection Features — GDPR, 23 Languages & More

PrivacyPromptAI supports 23 European languages: English, German, French, Spanish, Italian, Portuguese, Dutch, Swedish, Danish, Finnish, Norwegian, Polish, Czech, Greek, Hungarian, Romanian, Bulgarian, Croatian, Slovak, Slovenian, Estonian, Latvian and Lithuanian. Detection adapts to language-specific name structures, phone formats, postal codes and national ID patterns. See the full Language Support page →

Yes — XLSX and CSV file support is available on the Pro plan ($20 one-time) →. Spreadsheets are processed via SheetJS entirely in your browser. Each sheet is extracted as structured text and passed to the PII detection engine. Hidden rows and filtered data are also processed. Free users can copy spreadsheet cell content and paste it into the homepage tool →

Yes — custom local keywords are a Pro feature. Once activated, you can enter any term (a project codename, internal reference, company name) in Settings → and it will be detected and redacted automatically in every document you process. Keywords are stored only in your browser's localStorage and are never sent anywhere.

Yes. Once the page has fully loaded, all PII detection and redaction runs as JavaScript in your browser — no internet connection is required. You can load the page, disconnect from Wi-Fi, and the tool continues to work perfectly. You can also install PrivacyPromptAI as a Progressive Web App (PWA) for persistent offline access from your home screen.

No. PrivacyPromptAI keeps no logs, history, or records of any redaction you perform. When you close or refresh the tab, the text is gone. Nothing is written to our servers because there are no servers involved — the entire tool runs inside your browser's JavaScript engine. Your Settings page allows you to wipe all local browser data including any stored preferences at any time. If you need to keep a copy of the redacted output, pCloud offers Swiss encrypted cloud storage with client-side encryption — nothing is readable even to the provider.

Yes. In Settings you can toggle entity recognition (names and organisations) and credit card / financial pattern detection on or off independently. The remaining PII types — emails, phone numbers, IBANs, IP addresses, dates, and national IDs — are always active as part of the core detection engine. Pro users can additionally define custom keywords to be detected and redacted automatically.

Pattern-based detection (emails, IBANs, phone numbers, credit cards, IP addresses) is highly accurate — false negatives are rare. Named Entity Recognition for personal names can occasionally miss unusual or hyphenated names, or produce false positives for common words used as names in context. For high-stakes documents — legal, medical, HR — always review the redacted output before sharing. Pro users can add custom keywords to catch any domain-specific terms the engine may not recognise. See the FAQ for more detail on how detection works.