QR Code Reader

Decoding a QR code from an image you already have is a different situation from pointing a phone at a poster, and it has one significant advantage: you get to see what the code contains before anything acts on it.

Check the destination before you follow it

A QR code is just encoded text, and most of the ones in the wild encode a web address. Nothing about the visual appearance of a code tells you where it points, which is precisely what makes malicious codes effective. Stickers placed over legitimate codes on parking meters, restaurant tables and payment terminals are a well-documented tactic, because the substitution is invisible to the person scanning.

Reading a code here shows you the decoded text without opening it, which is the safe order of operations. Worth looking for: a domain that doesn't match the organization you expect, a link shortener hiding the real destination, lookalike spelling, and any code that resolves to a login or payment page you did not initiate. A physical sticker sitting on top of printed artwork is itself a warning sign.

What else codes can contain

Not every code is a link. The format supports plain text, contact cards, calendar events, geographic coordinates, phone numbers, and network credentials. A Wi-Fi code hands your device a network name and password, and an event code writes to your calendar, so seeing the decoded content first is useful well beyond phishing.

Why a decode fails

When a code won't read, the cause is usually recoverable. Insufficient resolution is the most common: a code photographed from far away may not have enough pixels per module for the boundaries to be distinguished. Motion blur and poor focus have the same effect. Glare from a glossy surface or a screen can wash out an entire region.

Missing margin is the failure that surprises people, because the code itself looks perfect. A code cropped tightly to its edges, or screenshotted without the surrounding white space, loses the quiet zone the decoder needs to identify its boundaries. Re-cropping with generous space around the code fixes a surprising number of apparently unreadable images. Extreme angles and low contrast, particularly light-on-dark color schemes, account for most of the rest.