PDFLean

PDF PII Detection

Cloud Processing (Encrypted)

Automatically detect sensitive entities in a PDF — emails, national IDs, phone numbers, bank cards, SSNs, IPs, names, and addresses — and export an audit report with per-page attribution. Detection is performed by the server-side Presidio NER engine (free account required, cloud-quota billed); text extraction in your browser is input preparation only.

Cloud processing: encrypted in transit, results kept for 24 hours then automatically deleted, never used for AI training · ≤ 50 MB per fileThis cloud tool needs a free account · Local tools: free & unlimited, no sign-up

How this tool works (free account required)

  1. Upload your file above. Each tool shows whether it runs Local (in your browser) or Cloud (encrypted servers).
  2. Adjust options, then click the Run / Convert / Process button.
  3. Wait for progress, then download the result. Use Next tools to continue.

If buttons are not visible, enable JavaScript — the interactive workspace loads in your browser.

pdflean://pdf-pii-detectSECURE CLOUD

Cloud NER detectionSensitive entities are detected by the server-side Presidio NER engine (free account required, billed as one cloud task per run); in-browser text extraction is input preparation only.

Drag files here, or click to browse

Supports application/pdf

Cloud mode: this file is uploaded over an encrypted connection and processed on our server. Files are automatically deleted after 24 hours and never used for AI training.

What is it?PDF PII Detection is a PDFLean security & privacy tool: Automatically detect sensitive entities in a PDF (emails, national IDs, phone numbers, bank cards, SSNs, IPs, names, and addresses) and export an audit report with per-page attribution.

TL;DR: Automatically detect sensitive entities in PDFs — emails, national IDs, phone numbers, bank cards, SSNs, IPs, names, and addresses — and export an audit report with per-page attribution and confidence scores. Detection runs on the server-side Presidio NER engine (requires a free account, cloud-quota billed, encrypted transport, deleted within 24 hours, never used for AI training); in-browser text extraction is input preparation only. Only text layers are scanned (scans need OCR first).

About this tool

Automatically detect sensitive entities in PDF files — emails, national IDs, phone numbers, bank cards, SSNs, IPs, names, and addresses — with an exportable audit report in JSON. Detection is performed by the server-side Presidio NER engine: it requires a free account, counts against your plan quota, and recognized entities are mapped back to their pages in the report. Text is extracted from the PDF in your browser as input preparation (no result is produced locally); only the extracted text is sent over an encrypted channel and auto-deleted within 24 hours. Files are never used for AI training.

How to use

1

Upload PDF(s) with a text layer (multi-file batch supported)

2

Click Scan — the server-side NER engine performs the detection (one cloud task per run)

3

Review the per-type summary and per-hit list (with page numbers and confidence), or export the audit JSON

FAQ

Is my file safe?

Detection is a cloud process: the extracted text travels over an encrypted channel to the server-side NER engine, is deleted within 24 hours, and is never used for AI training. The PDF itself produces no result file in your browser — text extraction is input preparation only.

Are there any limits?

Detection runs on the server-side Presidio NER engine and requires a free account; each run counts as one cloud task even for multi-file batches. Combined text beyond 200,000 characters is scanned only up to the cap; only text layers are recognized (scans need OCR first); JSON audit report export is supported.

What is the confidence score in the report?

The server-side NER engine assigns each detected entity a 0-1 confidence score — the higher, the more certain the model. The report summarizes by type and lists each hit with its page number and confidence so you can double-check.

What if names or addresses are not detected?

The server-side NER covers named entities such as persons and addresses; if your document is a scan, run OCR first to generate a text layer. Pattern-based formats (national IDs, bank cards, ...) are validated by the same server-side models.

Can I process multiple files in bulk?

Yes. This tool natively supports selecting multiple PDF files at once for batch PII detection (one cloud task per run) and exporting an aggregated audit JSON report. If you need to chain detection with other operations (such as watermarking or conversion), you can also use the Batch Processing workflow.

Is the processed file still editable?

This tool only detects sensitive entities and exports an audit report (or JSON) without modifying your original PDF. Your original file remains untouched and fully editable.