yescord

Take a screenshot and black out what is private

Capture, cover names and tokens, save a PNG. The image never leaves this tab.

or paste one (Ctrl+V, or ⌘V on a Mac), or drop it on this page. Pasted and opened images are redrawn onto a canvas, so their EXIF data (camera, GPS position, date) does not carry into the saved PNG.

Why redact before you share

A screenshot pasted into a chat, a ticket or an AI assistant often shows more than the question needs: the name and e-mail in a mail header, a customer's address, an API key in a terminal, the other tabs in the tab strip. Once the image is uploaded, the service has all of it, and deleting the message later does not tell you who already looked. Covering the private parts here, before the image goes anywhere, is the only point at which nobody else has seen them. An AI chat cannot do this step for you, because it only receives the image after the upload.

How to use it

  • Get an image in. “Capture screen, window or tab” opens the browser's own picker. Choose what to share; the page grabs one frame at the full size the browser reports (a 1920×1080 screen gives a 1920×1080 image) and stops the capture immediately, so the browser's “sharing” bar disappears. Or paste an image you already copied (Print Screen, Win+Shift+S, Ctrl+Shift+⌘+4 on a Mac, or “Copy image” in a browser), drop an image file on the page, or open one.
  • Redact. With “Solid box”, drag over what must go: the rectangle is painted in one flat colour (black unless you pick another). “Pixelate” turns the dragged area into large blocks, each one the average colour of the pixels it covers. Blocks are about 1/60 of the image's longer side and never smaller than 16 pixels: 32 pixels on a 1920-wide screenshot.
  • Mark up. Rectangle draws an outline, Arrow draws a line with a head at the end where you let go, Text puts the typed label where you tap. Crop: drag over the part to keep, then press “Apply crop”. Each step can be undone; “Start over” throws the image away.
  • Save or copy. “Save PNG” downloads yescord-screenshot-<date>-<time>.png at the image's full resolution; the line under it lists the PNG chunks written, so you can see there is no text or EXIF block in it. “Copy” puts the same PNG on the clipboard, ready to paste into a chat.

Solid boxes, coarse blocks, and why not blur

A solid box replaces every pixel with the same colour, so nothing of the original survives in the file. Blur and fine pixelation keep a trace of every letter: each blurred or averaged pixel still depends on the text below it. When the font and the kind of content are known (an e-mail address in a mail client, a 6-digit code), an attacker can render every candidate, blur or pixelate it the same way, and pick the one that matches. That has been used to recover redacted text from real screenshots; the “Depix” tool published in 2020 does it for pixelated passwords. Coarse blocks help because one 16-pixel block swallows most of a letter at normal screen text sizes, but they are still a weaker cover than a solid box. For passwords, keys and account numbers, use the solid box.

Redaction here is burned into the pixels. The saved PNG has one layer; there is no hidden original under the box, unlike a shape drawn over text in some document editors.

What it does not do

  • It does not find private text for you. You choose every area; look over the whole image, including tab titles, bookmarks and notification pop-ups, before you save.
  • It is not a scrolling or full-page capture. It takes what is on screen in that one frame.
  • It keeps nothing. There is no upload, no account and no browser storage; closing the tab forgets the image. The only request after the page loads is an anonymous, cookie-free page-view count (GoatCounter), which never sees the image.
  • Screen capture needs a desktop browser: Chrome from version 72, Edge from 79, Firefox from 66 and Safari from 13 (MDN, checked 2026-09-26). Phones and tablets cannot capture the screen from a web page; take the screenshot with the device's buttons and open it here.
  • “Copy” needs a browser that lets pages write images to the clipboard (Chrome 76, Edge 79, Safari 13.1, Firefox 127). Where it cannot, the page says so; use Save PNG instead.
  • The saved image is always a PNG. A photo saved as PNG is several times larger than the JPEG it came from; for screenshots of text, PNG is usually the smaller and sharper choice.

What has actually been tested

Headless Chrome on Linux, driven by the automated test in this site's repository on 2026-09-26: a capture with Chromium's fake screen source came out at exactly the track's reported size, with every capture track ended afterwards; a black box saved into a PNG whose covered pixels were all identical when decoded; the saved PNG held only IHDR, IDAT and IEND chunks, also when the input was a JPEG carrying an EXIF block; and no request left the page after it had loaded. Firefox, Safari and phones have not been tested.