Document redaction

Redact personally identifiable information (PII) from documents (pdf, png, jpg), Word files (docx), or tabular data (xlsx/csv/parquet). Please see the User Guide for a full walkthrough of all the features in the app.

To extract text from documents, the 'Local' options are PikePDF for PDFs with selectable text, and OCR with Tesseract. Use AWS Textract to extract more complex elements e.g. handwriting, signatures, or unclear text. For PII identification, 'Local' (based on spaCy) gives good results if you are looking for common names or terms, or a custom list of terms to redact (see Redaction settings). AWS Comprehend gives better results at a small cost.

Additional options on the 'Redaction settings' include, the type of information to redact (e.g. people, places), custom terms to include/ exclude from redaction, fuzzy matching, language settings, and whole page redaction. After redaction is complete, you can view and modify suggested redactions on the 'Review redactions' tab to quickly create a final redacted document.

NOTE: The app is not 100% accurate, and it will miss some personal information. It is essential that all outputs are reviewed by a human before using the final outputs.

Try an example - Click on an example below and then the 'Extract text and redact document' button:

Examples

Choose a PDF document or image file (PDF, JPG, PNG)

Choose text extraction method. Local options are lower quality but cost nothing - they may be worth a try if you are willing to spend some time reviewing outputs. If shown,AWS Textract has a cost per page - £1.14 ($1.50) without signature detection (default), £2.66 ($3.50) per 1,000 pages with signature detection. Change this in the tab below (AWS Textract signature detection).

Local model - selectable text Local OCR model - PDFs without selectable text

Choose a local OCR model. "tesseract" is the default and will work for documents with clear typed text. "vlm" will call the chosen vision model (VLM) to return a structured json output that is then parsed into word-level bounding boxes. "hybrid-vlm" is a combination of Tesseract for OCR, and a second pass with the chosen vision model (VLM). "hybrid-paddle-vlm" is a combination of PaddleOCR with the chosen VLM.

tesseract hybrid-vlm vlm

Choose personal information detection method. The local model is lower quality but costs nothing - it may be worth a try if you are willing to spend some time reviewing outputs, or if you are only interested in searching for custom search terms (see Redaction settings - custom deny list). If shown, AWS Comprehend has a cost of around £0.0075 ($0.01) per 10,000 characters.

Only extract text (no redaction) Local

If you only want to redact certain pages, or certain entities (e.g. just email addresses, or a custom list of terms), please go to the Redaction Settings tab.

Output summary

Output files

Choose a Word or tabular data file (xlsx or csv) to redact. Note that when redacting complex Word files with e.g. images, some content/formatting will be removed, and it may not attempt to redact headers. You may prefer to convert the doc file to PDF in Word, and then run it through the first tab of this app (Print to PDF in print settings). Alternatively, an xlsx file output is provided when redacting docx files directly to allow for copying and pasting outputs back into the original document if preferred.

Try an example - Click on an example below and then the 'Redact text/data files' button for redaction, or the 'Find duplicate cells/rows' button for duplicate detection:

Examples

Choose Excel or csv files

Choose PII detection method. AWS Comprehend has a cost of approximately $0.01 per 10,000 characters.

Local

Output result

Output files

Document redaction

Try an example - Click on an example below and then the 'Extract text and redact document' button:

Try an example - Click on an example below and then the 'Identify duplicate pages/subdocuments' button:

Matching Strategy

Analysis summary

Full Text Preview of Selected Match

Downloadable Files

Try an example - Click on an example below and then the 'Redact text/data files' button for redaction, or the 'Find duplicate cells/rows' button for duplicate detection:

Duplicate Analysis Results

Remove Duplicate Rows