Halaman ini masih berbahasa Inggris.
Redact
Find and remove Indonesian PII, storing only offsets and never the values.
Jalaran Redact finds and removes personally identifiable information from text, tuned for Indonesian formats such as NIK and NPWP, and stores only offsets, never the values it found.
Who Redact is for
For anyone under UU PDP handling documents that contain personal data and must be shared, published or processed. Also for developers preparing production data for a test environment, which is where a lot of breaches quietly begin.
What Redact does
Redact combines deterministic pattern matching for structured Indonesian identifiers (NIK, NPWP, phone numbers) with optional AI assistance for names and addresses, which have no fixed format. The design decision worth knowing is that only character offsets are stored, never the detected values, so using the tool does not create a second copy of exactly the data you were trying to protect. A redaction tool that logged what it found would be a liability rather than a control.
- Indonesian NIK, NPWP, phone and address patterns
- Deterministic patterns plus optional AI assist
- Findings stored as offsets, never as separate copies of the values
- Runs on self-hosted inference
How Redact works
Paste the text
A document, a dataset extract, anything containing personal data.
Run the deterministic patterns
NIK, NPWP, phone numbers and other structured Indonesian identifiers, matched by format.
Add the AI pass if needed
Names and addresses have no fixed format, so pattern matching alone cannot find them.
Review before relying on it
No detector finds everything. Read the output: this is the step that must not be skipped.
What Redact does not do
No PII detector finds everything, and this one does not either: unusual formats, names as ordinary words, and personal information implied rather than stated will be missed. Always review the output before treating text as safe to share; the tool reduces the work, it does not discharge the obligation. It handles text and not images, PDFs with scanned content, or structured databases, and it is tuned for Indonesian formats.
Common questions
- Will it find all the personal data?
- No, and treating it as if it does is the dangerous mistake. Unusual formats and information implied rather than stated get missed. Review the output before sharing anything: the tool reduces the work, it does not remove the responsibility.
- Does using it create another copy of my sensitive data?
- Yes, one: the document you paste is kept in your account, with its redacted version, until you delete it. The findings are stored as character offsets only, never as separate copies of the detected values.
- Does it work on PDFs and images?
- No. It processes text. Scanned documents and images are outside its scope.