Auto-Redact PDF Online — AI PII Detection & Removal

The AI detects: full names, Social Security numbers (SSNs), email addresses, phone numbers, physical addresses, dates of birth, credit card numbers, passport numbers, driver's license numbers, and bank account numbers. Detection is context-aware, not just pattern-matching, so it catches PII that regex tools miss.

Yes. Unlike tools that only add a visual overlay, WhiskFile's redaction permanently removes the underlying text and replaces it with black rectangles rendered into the page image. The redacted content cannot be recovered by highlighting, copying, or editing the PDF.

Yes. Before uploading, toggle which PII categories to detect — for example, enable only SSNs and credit card numbers while keeping names visible. You can also add custom words or phrases to redact.

Yes. Your PDF text is extracted and sent to the AI model for PII detection in a single request. The content is not stored, logged, or used for AI training. The file is deleted from our servers within 1 hour of processing.

Manual redaction (like Smallpdf and PDF2Go) requires you to find and highlight each piece of sensitive text yourself. WhiskFile's AI automatically scans every word in the PDF, identifies PII by context, and redacts it — no manual selection needed. This is faster and catches PII you might miss.

Regex tools only match fixed patterns (e.g. ###-##-####). Claude AI understands context — it detects names, addresses, and identifiers in varied formats, abbreviations, and natural language that patterns miss.

Yes. In the Advanced options, type comma-separated words or phrases. These are redacted alongside AI-detected PII — useful for company names, project codes, or any sensitive terms specific to your document.

Black (default), white, and red. Black boxes are the standard for legal and compliance documents. White blends with the page background. Red is useful for marking redactions that need review.

Yes. WhiskFile's PII categories map directly to GDPR personal data definitions: names, addresses, emails, phone numbers, dates of birth, and identification numbers. Use this tool to de-identify documents before archiving or responding to data subject requests.

Yes. The tool can redact patient names, dates of birth, addresses, phone numbers, SSNs, and other HIPAA-defined identifiers. For Safe Harbor de-identification, enable all 9 PII categories to remove all 18 HIPAA identifiers found in the document.

No. WhiskFile rasterizes each page as a 200 DPI image and draws opaque rectangles over PII. The underlying text is permanently removed — it cannot be recovered by highlighting, copying, searching, or editing the PDF.

Typically 15-30 seconds depending on document length. The AI scans all extracted text in a single request, then redaction boxes are drawn on each page image.

Yes, if the scanned PDF has an OCR text layer. The AI works on extracted text, so PDFs with selectable text (native or OCR) are supported. For image-only scans without a text layer, run OCR first using WhiskFile's OCR PDF tool.

Yes. Auto-Redact PDF on WhiskFile is completely free with no account required. Free users get 7 tasks per day with files up to 50 MB. There are no watermarks on any plan. Pro plans ($4.99/month) unlock unlimited tasks, 500 MB files, batch processing, and API access.

Yes. All file transfers use TLS 1.2+ encryption. Files are processed in isolated containers, scanned for malware with ClamAV, and automatically deleted within 1 hour. WhiskFile does not use your files for AI training or share them with third parties.

Yes. Auto-Redact PDF works on any device with a modern browser — iPhone, Android, iPad, tablets, Chromebooks, and desktop computers. No app download required. The interface adapts to your screen size automatically.

Auto-Redact PDF accepts PDF files as input and produces PDF output. The maximum file size is 50 MB on the free plan (500 MB for Pro users). There are no page count limits.