Anonymization
John Carter [PERSON]
In a PDF: a black box, with no text left behind it.
Sharing a document with a third party. Irreversible.
How it works
How data anonymization works with NymData
NymData finds the PII and other sensitive data in your documents and databases, then gives you an anonymized copy, with a report that proves what was done.
How do you anonymize documents automatically?
With NymData, you select a file, a folder or a database.
The PII detection engine, installed on your own server, finds names, addresses, emails, bank account numbers and other sensitive data, then replaces them using the method you choose.
You get an anonymized copy, the source file is never changed, and an audit report documents the processing.
In 4 steps
From raw data to an anonymized copy
- Step 1 · Source
You select a source
A file, a folder with all its subfolders, or a PostgreSQL or MySQL database.
Formats can be mixed: PDF, Word, Excel, PowerPoint, CSV and text files are processed in the same project.
You can exclude folders or file types (archives, temporary files…).
- Step 2 · Detection
NymData detects the sensitive data
The detection engine combines AI and rules, and runs on your server.
Each detection gets a confidence level: high, medium or low.
To check before you act, Mapping mode scans your data without changing anything: you see what was found, by data type and by file.
- Step 3 · Method
You choose the method
Replace the data with its type ([PERSON]), with realistic fake names, with a constant identifier (the same one for the same person), or choose one method per data type.
An allowlist protects terms that must never be touched, such as your organization's name.
- Step 4 · Result
You get the copy and the report
The anonymized copy is created in a separate folder, with the same structure and the same layout.
The source file is never changed.
An audit report documents the processing: date, data types detected, method, volumes.
Detection
The sensitive data NymData recognizes
- Identity
- Full names
- Dates of birth
- French social security numbers (NIR)
- Passports
- ID cards
- Government IDs
- Contact details
- Addresses
- ZIP and postal codes
- Emails
- Phone numbers
- Financial
- IBANs
- Credit cards
- French company IDs (SIRET)
- US bank account and routing numbers
- Financial identifiers
- Organizations
- Technical
- IP addresses
- Network identifiers
- Passwords and secrets
- License plates
- Dates
- Sensitive text
Have an identifier specific to your business (case number, contract number, patient ID)? Custom detection rules can be added as an option.
Methods
Which method should you choose?
It depends on what you need the data for: sharing, analytics or testing.
Realistic pseudonymization
John Carter Michael Brooks
Testing, demos, documents that need to stay readable.
Constant identifier
John Carter PERSON_7b94f21c
Analytics and AI: the same person keeps the same identifier across the whole project.
Custom
GB82 WEST 1234 5698… [IBAN]
John Carter Michael Brooks
One method per data type: here, IBANs are replaced with [IBAN] and names are replaced with realistic fake names.
Under the GDPR, pseudonymized data is still personal data. Pseudonymization vs anonymization
Scheduled processing
Anonymize your historical data, then every new file automatically
A project can run on its own: daily, weekly, monthly or on your own custom schedule.
On each run, NymData only processes what is new or has changed.
- Historical data: a first run processes years of archives, later runs only pick up new files.
- Shared folders: documents dropped into a folder are anonymized on the next run.
- Databases: feed a test or analytics database with anonymized data, picking up only the rows changed since the last run.
- Retention periods: anonymize records that reach the end of their retention period instead of deleting them.
Each scheduled run produces its own audit report.
Formats
What NymData does with each format
Each format is handled its own way, so the anonymized file stays as usable as the original.
- Redact a PDF →
PDF
Detected data is redacted and truly removed from the file: it cannot be found by searching or copying. The author and other hidden file information are erased.
- Anonymize Word and PowerPoint →
Word and PowerPoint
Data is replaced directly in the document, including in tables, headers, footers and speaker notes. The formatting stays the same.
- Anonymize Excel and CSV →
Excel and CSV
Only the sensitive cells change: columns and sheets stay in place, and a CSV file keeps its format so it can be imported again.
Text files
Sensitive data is replaced, and the rest of the text is kept as is.
- Anonymize a database →
PostgreSQL and MySQL
NymData reads your database and writes an anonymized copy to another database: the source database is never changed.
Audit report
Every processing run is documented
Timestamped and sealed with a SHA-256 hash, the report serves as evidence for your privacy officer (DPO), your auditors or your customers:
- Date and time of processing
- Data types detected
- Method applied
- Number of items processed
- SHA-256 hash of the report
Sample
Audit report AUD-20260729-0037
- Status
- Completed
- Documents scanned
- 6
- Sensitive data detected
- 20,359
- Data anonymized
- 19,217
- High confidence
- 90%
e964db7dd4de74f80b05930f01370674198033c8eab627f5946adafe5e40c51a
FAQ
Frequently asked questions about how NymData works
Can I anonymize an entire folder?
Yes. NymData processes a folder and all its subfolders in a single operation, whatever the mix of formats, and creates an anonymized copy of each document.
Can new files be anonymized automatically?
Yes. A project can be scheduled daily, weekly, monthly or on a custom schedule.
On each run, only new or changed files are processed.
Are the original files changed?
No. The source file is never changed: NymData creates an anonymized copy in a separate folder, on your servers.
NymData does not store the content of your documents.
Is anonymization reversible?
The Anonymization method is irreversible: the data is replaced by its type, or covered by an opaque box in a PDF.
The pseudonymization methods produce consistent replacement values within a project.
Under the GDPR, pseudonymized data is still personal data.
Can I review detections before anonymizing?
Yes. Mapping mode scans your data without changing anything: you review what was found, by data type and confidence level, before you run the anonymization.
Do you need technical skills to use NymData?
No. NymData runs in a web browser: you create a project, choose the source and the method, then start the processing.
Installation on the server is done once, with your IT team.
Watch a full folder get anonymized in front of you
Live demo, online or on-site.