Privacy-First Data Anonymization

Anonymize your research data
before it leaves your machine.

Most AI tools log what you upload. DataMask anonymizes your datasets entirely in the browser — no server, no backend, no data transmitted — so you can safely use AI tools without exposing real values.

100% local No upload No backend Open source · MIT
Why DataMask

Everything runs in your browser

No account required. No installation. Open the file and start anonymizing — your data never touches a server.

🔒

100% Local Processing

Every transformation — noise injection, sigmoid distortion, SHA-256 hashing — happens inside the browser's JavaScript engine. There is no API call, no server log, no data retention policy to worry about.

⚡

Three Protection Tiers

Choose the level of anonymization that fits your workflow.

Lite — reversible Strong — sigmoid distortion Vault — SHA-256 cryptographic
🏷️

Text Label Substitution

Sample names, strain IDs, and group labels are replaced with consistent tokens (SAMPLE_001, SAMPLE_002, …). The same original value always maps to the same label. Export the mapping for full traceability.

📋

Flexible Export

Download anonymized data as .xlsx, .csv, .tsv, .txt, or .json. Copy directly to clipboard for instant paste into a chat. Custom filename prompt on every download.

Workflow

Four steps. No setup.

From raw dataset to anonymized export in under a minute.

STEP 01

Upload

Drag & drop or paste CSV from clipboard. Supports .xlsx, .xls, .csv, and .txt.

STEP 02

Select Columns

Check the columns to anonymize. Types are auto-detected. Pick numeric, text, or both.

STEP 03

Choose Mode

Lite, Strong, or Vault. Adjust noise level, decimal places, and shape-preservation options.

STEP 04

Download

Export in your preferred format. Optionally download the original→label mapping.

Open Source

Read the entire source in 5 minutes.

DataMask is a single self-contained HTML file — no bundler, no framework, no hidden dependencies. Every transformation is visible in plain JavaScript. Fork it, embed it, or audit it.

MIT License · github.com/mbaffour/datamask