Screenshot Redaction Tool

Redact PNG, JPEG, or WebP screenshots entirely in the browser. Draw opaque regions, add source-pixel padding, inspect exact coordinates, review the newly encoded PNG, and optionally download a private redaction manifest.

Choose a screenshotPNG, JPEG, or WebP · up to 20 MB and 25 megapixels

The screenshot, redactions, encoded PNG, and manifest remain in this browser. Nothing is uploaded.

Choose a screenshot to begin.

Draw opaque regions over names, account details, messages, tabs, notifications, or other private content.

How to Redact a Screenshot

Choose a PNG, JPEG, or WebP screenshot.

Drag over every private area. Use 100% or 200% zoom to inspect small text and edges.

Add source-pixel padding when a tight rectangle could leave anti-aliased text, shadows, or narrow borders visible.

Prepare the encoded release preview, inspect the actual PNG at source size, and download it only after completing a full-frame review.

Opaque Replacement Versus a Removable Overlay

Placing a black annotation above content can be unsafe when the original layer, object, or document remains present underneath.

This tool draws the decoded screenshot onto a new canvas and then paints every committed rectangle with fully opaque black pixels before PNG encoding.

The selected coordinates in the newly encoded PNG therefore contain black output pixels rather than the original screenshot pixels.

This applies only to the downloaded redacted PNG. It does not modify or erase the source file.

Why the Encoded Preview Is Separate

The editor displays white borders and number labels so committed regions can be reviewed and removed.

An unfinished drag also appears as a temporary translucent preview.

None of those editor guides are exported. The release preview is built on a separate canvas containing only the decoded image and committed opaque regions.

The displayed release image is loaded from the generated PNG Blob rather than reused from the editor canvas.

Conservative Source-Pixel Boundaries

A screenshot may be displayed smaller or larger than its source dimensions, causing pointer positions to map to fractional source pixels.

The calculator floors the absolute left and top edges and ceils the absolute right and bottom edges.

This avoids the thin far-edge leak that can occur when a fractional starting coordinate is rounded separately from the rectangle width.

Optional padding expands future regions on every available side and clips the result to the image.

What to Review Before Sharing

Privacy review should include more than names and email addresses.

Inspect avatars, faces, maps, account IDs, tabs, notification previews, filenames, device details, internal URLs, customer records, and identifying conversation context.

Review the entire encoded output at source size rather than relying only on the fit-to-width editor view.

Screenshot privacy review checklist

Inspect every part of the encoded release image, including direct identifiers and identifying context.

Screenshot privacy review checklist
Category	Examples	Where to inspect
Identity	Names, faces, avatars, signatures, usernames	Headers, comments, profiles, reflections, and thumbnails
Contact and accounts	Email, phone, customer IDs, order numbers, wallet addresses	Sidebars, footers, autofill, forms, and notification banners
Location and devices	Addresses, maps, coordinates, Wi-Fi names, clocks, device names	Status bars, map previews, filenames, and system panels
Business and security	API keys, tokens, internal URLs, tickets, revenue, customer data	Tabs, consoles, network panels, tooltips, and background windows
Identifying context	Projects, dates, conversations, relationships, or unique events	Review the whole frame, not only the most obvious text

Swipe horizontally to view the full table.

Output Filename and Metadata

The original filename may reveal a person's name, project, date, account, ticket, or case reference.

The tool therefore defaults to the generic filename redacted-screenshot.png.

Canvas export creates a newly encoded PNG from the bitmap rather than copying the source file container.

Do not describe the output as containing no metadata of any kind. The browser may add new encoder-generated information such as standard resolution metadata.

The Private Redaction Manifest

The optional manifest records source dimensions, normalized rectangles, the opaque replacement colour, and union pixel coverage.

It does not contain OCR text or copies of the covered pixels.

Coordinates can still reveal where information appeared and how much content was hidden, so the manifest should normally remain private.

Redaction Does Not Erase the Original

Downloading a redacted PNG creates another file.

It does not sanitize the original screenshot on a device, cloud drive, backup, messaging platform, clipboard, or recipient system.

Protect or delete other copies according to the sensitivity of the information and the rules that apply to the situation.

When This Tool Is Not Enough

Legal disclosure, litigation, classified information, medical records, formal subject-access responses, evidence handling, and regulated investigations may require approved software, documented procedures, trained review, and independent verification.

Use this browser tool as a controlled editing step rather than proof that a release is lawful, complete, or organisationally approved.

Privacy and Local Processing

Image decoding, region editing, coverage calculation, PNG encoding, preview creation, and manifest generation happen locally in the browser.

The screenshot and redaction information are not uploaded or permanently stored by the tool.

Redaction Rectangle Normalization

Fractional drag coordinates and optional padding are converted into conservative integer source-pixel boundaries and clipped to the image.

Formula variables

$x_{s}, y_{s}$: Starting source-pixel position
$x_{e}, y_{e}$: Ending source-pixel position
$p$: Additional source-pixel padding
$W, H$: Decoded image width and height
$R_{i}$: One committed redaction rectangle
$q$: Percentage of source pixels covered by the region union

Left edge

x_{0} = max (0, ⌊ min (x_{s}, x_{e}) - p ⌋)

Top edge

y_{0} = max (0, ⌊ min (y_{s}, y_{e}) - p ⌋)

Right edge

x_{1} = min (W, ⌈ max (x_{s}, x_{e}) + p ⌉)

Bottom edge

y_{1} = min (H, ⌈ max (y_{s}, y_{e}) + p ⌉)

Normalized region

R = (x_{0}, y_{0}, x_{1} - x_{0}, y_{1} - y_{0})

Union coverage

q = \frac{∣ ⋃ _{i} R _{i} ∣}{W H} \times 100%

Why does the tool use black rectangles instead of blur?

Blur and pixelation preserve transformed visual structure and may leave content recognizable. This tool uses fully opaque replacement pixels.

Does it automatically detect private information?

No. Names, faces, account details, messages, and every other region must be identified manually.

Why should I add edge padding?

Anti-aliased text, shadows, and narrow outlines may extend just beyond a tight selection. Padding creates a conservative border.

Can I redact using exact coordinates?

Yes. Enter source-image x, y, width, and height values. The current edge-padding setting is also applied.

Does PNG export remove all metadata?

It creates a new PNG from the decoded bitmap rather than copying source file chunks. The browser may still add new output metadata such as standard resolution information.

What does the private manifest contain?

It contains image dimensions, normalized rectangle coordinates, opaque colour information, and union coverage. It does not contain the covered pixels or OCR text.

Does the tool delete the original screenshot?

No. It only creates a new redacted PNG.

Are my screenshots uploaded?

No. Decoding, editing, encoding, previewing, and manifest generation happen locally in the browser.

References

Image Compressor

Compress JPG/JPEG, PNG, WebP, AVIF, HEIC, and HEIF locally with Target KB, per-image controls, before-and-after comparison, and ZIP downloads.

Image Converter

Convert static JPG, PNG, and WebP images locally with quality and transparency controls, before-and-after comparison, and ZIP downloads.

Image Resizer for Online Forms

Check and resize JPG, PNG or WebP images against online-form requirements for file size, pixels, format, aspect ratio and DPI.

Spin the Bottle

Spin an animated bottle to choose a random direction, or add 2–12 player names and let the bottle pick one person.