CSV & JSON PII Cleaner
Preview CSV, JSON, and NDJSON files in your browser, detect common sensitive patterns, then mask, hash, remove, or export cleaned data before sharing it with teams, forms, dashboards, or AI tools.
Drop a CSV, JSON, or NDJSON file
Use this for research data, analytics exports, contact lists, logs, and spreadsheet-like datasets. The file is read in this browser tab.
This tool detects common patterns only. It is not a legal, compliance, or forensic guarantee. Always review the preview and cleaned output before sharing sensitive data.
File summary
Detected columns
Review detected patterns and choose an action per column. The default action is applied when a column contains sensitive-looking values.
| Column | Detected patterns | Hits | Sample | Action |
|---|---|---|---|---|
| Analyze a file to see column findings. | ||||
Original preview
| No preview yet. |
Cleaned preview
| Cleaned preview appears after export. |
How to use this tool
- Drop a CSV, JSON array, object, or NDJSON file.
- Analyze the file to preview columns and sensitive-looking patterns.
- Choose mask, hash, remove, keep, or drop-column actions.
- Export the cleaned file and optional audit report.
Common use cases
Use this before sharing analytics exports, research datasets, CRM samples, spreadsheet extracts, log files, or AI prompt datasets. It is designed for practical cleanup before upload, review, or collaboration.
Privacy and limits
The selected file is processed in the current browser session. Large files can use significant memory, and browser performance varies. Pattern detection can miss unusual formats or flag false positives, so human review still matters.
FAQ
Does this upload my CSV or JSON file?
No server upload is required by this tool. It reads the selected file in the browser tab and creates downloadable outputs locally where supported.
What patterns can it detect?
The MVP checks common email, phone, URL, IPv4, card-like number, national ID-like, and token-like values. You should still review the output for context-specific sensitive data.
Is hashing reversible?
No. The built-in hash action creates a stable browser-side SHA-256 based label for matching repeated values, but it is not a substitute for a full data governance process.
Can it handle huge files?
This MVP is designed for practical browser cleanup and previews. Very large files depend on available memory and browser limits. A future worker/streaming lab mode can extend large-file handling.