Review before redacting
Detection is heuristic. A number can resemble an account, card or identifier without actually being sensitive. Review the match list before applying changes.
Privacy & Business Tools
Document Sensitive-Data Sanitizer
Detect common sensitive values in text, review every match, redact only what you approve, and export a safer copy before sharing. Processing stays in your browser.
1. Add text to sanitize
Paste text or import a .txt, .csv, .json, .log or .md file. Nothing is intentionally uploaded by this tool.
0 characters0 lines0 detected matches
2. Choose detectors
3. Review matches
0 of 0 matches selected for redaction.
Add text above to scan for sensitive values.
4. Create sanitized output
Important: automated detection can miss sensitive data and can flag harmless text. Review the complete output manually. For PDFs, Word files, images or legally required redaction, use a workflow that removes underlying content, metadata and hidden layers—not only visible text.
Useful before sharing
Use the sanitizer on support logs, sample payloads, copied emails, tickets, notes and test data before sending them to another person or system.
Not secure PDF redaction
This tool sanitizes extracted or pasted text. It does not prove that hidden layers, metadata, attachments or original document bytes have been securely removed.
Safer review workflow
- Paste or import the text that needs to be shared.
- Choose detector types and review every candidate match.
- Deselect false positives instead of redacting blindly.
- Choose the redaction style and create the sanitized copy.
- Compare the final copy with the original purpose and check again for missed sensitive context.
Common sanitization mistakes
- • Assuming pattern detection understands the meaning of every identifier.
- • Redacting visible text but sharing the untouched original file separately.
- • Forgetting names, addresses or business-specific identifiers that are not covered by a selected detector.
- • Treating masked text as anonymous when surrounding context can still identify a person.
- • Using a text sanitizer when the requirement is secure document redaction and metadata removal.
What should still be checked manually?
Contextual identifiers
Names, job titles, addresses, case details or combinations of ordinary facts may still identify someone even when common patterns are removed.
Secrets and tokens
Custom API keys, internal IDs and proprietary token formats vary. Review unusual strings even when no detector flags them.
Original document data
For PDFs, office files or images, secure redaction may also require removal of metadata, comments, hidden layers and embedded objects.
Frequently asked questions
Does detection guarantee that every sensitive value is found?
No. Detection is heuristic and can produce both false positives and false negatives. Review the original and sanitized output before sharing.
Is this secure PDF redaction?
No. This workflow sanitizes pasted or extracted text. It does not prove that hidden PDF layers, metadata, attachments, comments or original bytes were removed.
Why are some long numbers marked for review?
Long identifiers can resemble account, customer or reference numbers. The tool prefers review over silently deleting a value whose meaning it cannot know.
Are card-like numbers validated?
Card-like candidates can use a checksum check to reduce obvious false positives, but a passing pattern or checksum does not prove that the value is a real payment card.
Does the text leave my browser?
The sanitizer is designed to process the entered text locally in the browser. Your organisation may still have separate requirements for handling confidential or regulated data.
Related tools: Privacy Compliance Checklist · Privacy-First Password Generator · Document OCR.
Privacy note: processing is designed to happen locally in the browser. Even so, follow your organisation's handling, retention, legal and secure-redaction requirements for regulated or highly confidential information.