Image privacy redaction — Delivery workflow

Worked example

Share only the checked file named with a redact suffix, renamed to something neutral, and keep the unedited capture in a restricted folder.

Imagine a support team that sends customer screenshots to engineering. A workable routine is: capture, redact, reopen the saved file at 400%, rename it to something neutral such as ticket-4821-screen.png (the original name may include a person), and only then attach it. Keep the unedited capture in a restricted folder, never in the shared ticket.

The edit queue holds up to 12 images and 48 MB in total, so a batch of eight screenshots from one case can be redacted one after another. With two or more finished, a ZIP named redact-results.zip is available with numbered files; the results are kept up to 64 MB, then downloaded individually.

Keep the format in mind. A PNG screenshot stays a PNG; text-heavy captures usually look better and compress well as PNG, while a photo of a document is smaller as JPG at the cost of artifacts near the pixelated blocks. Do not re-compress a redacted file repeatedly, since that only adds noise.

The common mistake is sharing the whole working folder. Send only the checked export, and if a second person must confirm, give them the redacted file and ask them to mark anything legible. Note that you did redaction in a browser without uploading the file, which is helpful context for a security review.

Screenshot with a marked private region

Export with manually covered region

Hand over only the checked export, and keep the original in a restricted place for yourself.