How does the data anonymizer detect personal data?

It inspects each column header (matching names like email, phone, ssn, card, first_name) and also pattern-matches a sample of the values against email, phone, credit-card, IBAN, SSN, IP and date formats. Each field gets a suggested strategy that you can override.

Are the fake values realistic and consistent?

Yes. Emails look like real emails, cards are Luhn-valid 16-digit numbers, phones and SSNs follow real formats. Each original value is hashed to seed its fake, so the same input always maps to the same output — your joins, counts and duplicates stay intact.

Is my data uploaded anywhere?

No. The entire tool runs in your browser using JavaScript. Your file is parsed, masked and exported locally — there is no server, no network request and no logging of your data.

What masking strategies are available?

Keep as-is, full redaction, a stable hash token, fake full / first / last names, fake email, phone, credit card, IBAN, SSN, street address, city, company, username, IP address, and date shifting by a few days.

Can I use the output as test or demo data?

That is the main use. Because formats are preserved and values are internally consistent, anonymised CSV or JSON works as safe seed data for staging environments, demos, screenshots, bug reports and analytics without exposing real customers.

Does hashing let me re-identify people later?

The hash token is one-way and stable but not reversible, so you cannot recover the original value from it. It is useful when you need to preserve which rows share the same value (for grouping) without revealing what that value was.

Data Anonymizer — Gera Tools

The Data Anonymizer turns a spreadsheet or JSON export full of real personal information into a safe, shareable version in seconds — entirely inside your browser. Paste a CSV with a header row, or a JSON object or array of objects, and the tool scans every field for personally identifiable information (PII): email addresses, names, phone numbers, credit-card numbers, IBANs, social-security numbers, IP addresses, street addresses, cities and dates. It then proposes a sensible masking strategy for each column, which you can fine-tune before exporting a fully anonymised copy.

This solves a problem every developer, analyst and support engineer hits: you need realistic data to build a demo, file a bug report, populate a staging database or take a screenshot — but the only data you have contains real customers. Uploading that to an online “anonymizer” is itself a privacy risk. Here, nothing is ever sent anywhere; the parsing, masking and export all happen locally.

How it works

When you press Analyse, each column is classified two ways. First by its header name — a field called email, phone, ssn, credit_card or first_name is matched directly. Second by its values — a sample of cells is tested against format patterns, so a column of [email protected] style strings is recognised as email even if the header is cryptic. Each field is assigned a strategy you can override per-column or in bulk.

The fakes are format-preserving and deterministic. Credit-card replacements are Luhn-valid, phones and SSNs follow real shapes, and every original value is hashed to seed its replacement — so the same email always becomes the same fake email. That keeps row relationships, duplicate counts and joins intact, which naive random replacement would destroy. A live before/after preview highlights every changed cell in green; hover to see the original.

Example

Paste a customer export with columns full_name, email, phone, credit_card, city, signup_date. The tool auto-detects the first five as PII and leaves id and city choices to you. A row of Jane Doe, [email protected], +44 7700 900123, 4111 1111 1111 1111, London, 2023-04-12 becomes something like Amara Okafor, [email protected], +1 (415) 220-7781, 4539 8821 0067 1234, Lakeside, 2023-04-09 — same shapes, zero real data. Export the masked CSV or JSON, or copy it straight to your clipboard.

Every value is processed locally. Your real data is never uploaded, stored or logged.