Masking and restore — one complete, reversible loop
From batch processing to audited rollback, every step runs locally; every replacement is traceable to its coordinates and reversible to the original text.
Masking workflow
Batch folder selection with recursive scan; multi-file parallel processing (2 threads by default, configurable 1–8).
- Auto detection: column-keyword + content-sampling dual channel, merged; every result can be overridden in the confirmation dialog
- Results at a glance: per-file status, elapsed time, replacement count and output path
- Logs: timestamped entries you can copy, clear and export
- Ordered output: archived per task timestamp, original folder structure preserved
Mask + fake-data modes
Asterisk masking is the default; fake-data replacement keeps "one original value = one fake value" within a task, so statistics stay consistent.
- Custom masking template:
#keeps the original character in place; other characters are output as-is - Custom column rules: configure masking for any column you choose
Two restore channels
"Restore from mapping file" and "Restore from task history" work side by side.
- Position-level mapping: originals are written back by coordinates — file / sheet / row / column (paragraph and text-block positions in Word) — unaffected by masked-value collisions
- Pre-restore check: the processed file's hash is verified; if the file was altered manually, that file's restore is aborted with a warning
- Audit switch: on by default; tasks processed with auditing off keep no mapping and cannot be restored (clearly indicated in the app)
Rule management
Rules persist in a local SQLite database; add, edit, delete and toggle them in the UI.
- 7 built-in sensitive-field types, enabled out of the box
- Built-in rules allow parameter changes but cannot be deleted
- Mapping export: export the mapping workbook (xlsx) for restore and audit
7 built-in sensitive-field types
Checked and ready by default; combine them with custom column rules to cover your business-specific fields.
5 formats, kept as close to the original as possible
Excel keeps formats and styles; Word covers body text, tables, headers/footers and text boxes; CSV keeps its original encoding.
* Formula cells are skipped; only literal values are processed.
Privacy pledge: all local, zero telemetry
Every PrivaMask operation runs on your own computer: no network requests, no telemetry, no analytics, no accounts, no cookies. Source files are read-only; results go to the output folder you choose. See the Privacy Policy.