Back to public tools
All tools

Data Anonymizer

Pseudonymize and review CSV locally with column policies, suppression, date grouping, numeric buckets, masking, linkage controls, identifier scans and an audit report.

Choose how each CSV column should be kept, replaced, grouped, masked or removed. This is a local pseudonymization and review helper; it does not guarantee anonymity.

Source CSV

Header-based CSV only: 5 MiB input, 50,000 data records, 200 columns and 20 MiB output. Blank/duplicate header names are supported by column position. Scans use bounded email-, phone-, IPv4- and long-digit-like patterns; they can miss identifiers and flag ordinary text. Names and addresses need manual review.

How to use Data Anonymizer

Paste or import UTF-8 header-based CSV, inspect columns and choose per-column presets or actions. Suppress removes a field; token replaces values with fresh per-run placeholders. Date groups use year, quarter or month, and numeric buckets use exact [lower, upper) ranges. Masks retain a chosen Unicode prefix/suffix around [redacted], fully redacting short values. Shared token groups intentionally link matching values across columns; per-cell tokens remove repeated-value linkage.

Transform and review creates the complete CSV, source/residual scan summaries, rarity indicators and policy counts. Scans cover your selected columns; dates require canonical YYYY-MM-DD. Invalid dates or numbers can reject the run or be redacted. Choose up to four retained columns for combined rarity and a frequency threshold from 2 to 100. Review the six checklist items before deciding what to share; their state is included in the audit but does not block downloads. Limits: 5 MiB, 50,000 data records, 200 columns and 20 MiB output. Copy is capped at 2 MiB; downloads include every accepted cell. Formula-like text and headers are prefixed for spreadsheet export. No source, lookup map or draft is saved.

Example

For a contact export, suppress an unnecessary email column, tokenize member names, generalize birth dates to years and bucket an amount by tens. Scan retained notes for identifier-like candidates and review rare values or selected column combinations. Repeated-name tokens remain linkable within the chosen scope; this does not guarantee anonymity.

About Data Anonymizer

Pseudonymization, masking and generalization do not guarantee irreversible anonymization. Retained values, repeated patterns, row order and external information may still identify people. Identifier scans are heuristic and can miss names, addresses and other identifiers or flag ordinary text. Rarity counts and the self-reported checklist are review aids, not an identification probability or certification. Audit exports contain settings and record/column positions without raw cell values or token lookup maps; review metadata before sharing.

Processing stays in this browser tab without uploads or saved drafts. Reloading clears the workspace.

Stars and feedback

Loading…

Tool reference #039