Privacy Blur Flow

When you need to blur sensitive information in images before sharing them online, a focused privacy blur tool is safer than sending files through a full editor. Many people want to blur faces and license plates in photos, hide account numbers in screenshots, or redact text and email addresses in a single pass so they can safely post or send images without exposing private details.

This privacy blur flow is built for that job: it helps you hide sensitive parts of a photo, choose what area to keep, set JPEG quality for smaller file size, and export a redacted image where marked regions are replaced with solid color and noise instead of simple on-screen blur, making it harder to recover the original content.

Tool Overview

This workflow helps you hide sensitive parts of a photo, choose what area to keep, set JPEG quality, then save a new file. Most of the work runs in your web browser on your device. One optional step can ask a server to suggest boxes around sensitive areas.

Photos can show faces, text, or numbers you do not want to share. Cropping and solid redaction lower the risk that hidden detail can be recovered later. This tool combines blur-style marks on screen with solid fills and noise in the final file.

It fits anyone sharing screenshots or personal pictures online. Beginners can click through the steps. Careful users will check each box before they export.

Background & Concept Explanation

Redaction means replacing pixels so the old content is not shown. A blur on the editing screen is only a guide. The finished file uses flat color and random dots in those regions so the area is visibly removed.

Cropping cuts away outer parts of the picture. Here the crop is a rectangle in percent of the full image. Only the inside of that rectangle is kept in later steps.

JPEG quality sets how much compression you accept. Lower quality often means a smaller file and more visible loss.

Optional auto-detection sends image data out for analysis when you run it. The rest of the pipeline can stay on your machine.

Key Features

• Single image upload. You pick one file. It must be an image type and must pass the shared size limit.
• Redaction zones. You place rectangles on the picture in percent units. You can move them, resize from corners, or delete them. New clicks add a fixed-size manual box unless another box is active.
• Optional smart scan. A button reads the image as base64 and calls a backend task with the id privacy-blur-pro. Returned boxes are added to your list with types like face or text. On failure or empty results, you see a message and can still draw by hand.
• Minimum zones to continue. You need at least one redaction before you can go to the crop step.
• Crop rectangle. The default crop is centered at fifteen percent from the left and top with seventy percent width and height. You can drag the crop or resize from one corner. A safe margin toggle shows a padded border inside the keep area.
• Optimize preview. A hidden canvas draws the cropped image, fills overlapping redactions with gray plus small random dots, and can draw a margin stroke. The quality slider runs from zero point one to one in steps of zero point zero five.
• Scrub metadata toggle. The screen lets you turn metadata scrubbing on or off and describes GPS and device data. The live preview rebuilds when this changes.
• Final export. Another canvas pass builds the downloadable JPEG at your quality setting. The file name starts with privacy_redacted_ and ends with a time stamp.
• Download. You save the blob to disk from the last step.
• Start over. You can reset the whole flow and revoke blob links.

Common Use Cases

A user blurs account numbers in a screenshot, crops to the chat area, then exports a smaller JPEG.

Someone runs smart scan to get starter boxes, then tightens them by hand.

A person keeps safe margin on to remind themselves not to put text too close to the edge of the crop.

A helper exports at medium quality to email a smaller attachment.

How to Use This Tool (Step-by-Step)

1. Upload one image from the first screen or drop it onto the area.
2. On the redact step, add boxes or run smart scan. Adjust or remove boxes until you are satisfied.
3. Press continue. On the crop step, move or resize the keep area. Toggle safe margin if you want the guide.
4. Press confirm. On the optimize step, move the quality slider and tap the metadata option if you want to change it.
5. Review the live render preview. Press finalize export.
6. On the done step, wait for processing to finish, then download. Use start over if you need another image.

Calculations & Logic

Click-to-add box. A new manual redaction uses twenty percent width and height. Its top-left is placed near the click, shifted up and left by ten percent, then clamped so it stays inside the frame.

Crop to pixels. The tool multiplies crop percents by full image width and height to get source pixels for drawing.

Redaction overlap. For each redaction, the code converts percent to full-image pixels, shifts into crop space, and only fills where the box crosses the cropped region.

Noise loops. The optimize preview adds thirty random two-pixel dots per redaction area. The final export adds fifty such dots per area.

Safe margin stroke. When enabled, a stroke uses five percent of the smaller side of the cropped canvas.

Default quality. The session starts at zero point eight on the zero-to-one scale.

Reference Tables or Scales

Control	Range or default
JPEG quality slider	0.1 to 1, step 0.05, default 0.8
Initial crop (percent)	x 15, y 15, width 70, height 70
New manual redaction size	20 by 20 percent

Tips, Limitations, and Best Practices

Smart scan needs a network call and account access when the platform requires login. The footer text says processing stays in the browser; that is not fully true for that one optional action.

The export summary text claims zero bytes uploaded to servers. That line is only fully accurate if you never run smart scan.

The copy-link helper exists in code for the export screen but no button calls it, so only download is available in the UI.

The final JPEG comes from canvas encoding. Typical canvas output does not carry the same embedded metadata as a camera file; the scrub toggle mainly affects the optimize preview refresh in this folder.

If you need at least one redaction and you only want to crop, add a tiny box you will crop out, or place a box outside the final crop area so it does not affect the saved pixels.

Reset clears blob URLs and returns you to upload.

The optimize preview only watches the number of redaction zones, not every drag. Add or remove a zone if the preview looks stale after you move boxes.