Secrets and keys
API keys, access tokens, JWTs, bearer tokens, .env values, URL query tokens, and high-entropy strings.
Detector coverage
What Redaktr looks for before you share a screenshot.
Redaktr combines OCR text matching, visual detection, custom policy terms, and manual redaction so risky screenshots are easier to review before they reach AI or chat.
Built-in categories
Personal data
Email addresses, phone numbers, faces, and card-like numbers validated with a Luhn check.
Business data
Account IDs, customer IDs, IP addresses, internal labels, customer names, domains, and codenames.
Visual codes
QR codes and barcodes that may carry links, contact details, or tokens that are easy to miss visually.
Policy terms
Custom terms, regexes, allowlists, and shareable .redaktrpolicy files for team-specific confidential words.
Manual redaction
Draw a box over anything the detector missed or anything you simply do not want visible.
How findings are handled
| State | Meaning | Expected action |
|---|---|---|
| Redact | High-confidence sensitive content that should be covered in the cleaned output. | Review the box and keep it redacted unless it is clearly safe. |
| Check | Potentially sensitive content that needs judgment, such as an account ID or configured term. | Choose whether to redact it or leave it visible. |
| Visible | A finding you explicitly decided can remain in the screenshot. | Confirm it is safe before copying or exporting. |
Limits are part of the workflow
No screenshot redaction detector catches everything automatically. Redaktr narrows the review surface, explains likely leaks, and keeps manual control in the loop.