How Do QR Codes Work?

Every piece of information a QR code holds, a website address, a payment reference, a Wi-Fi password, has been converted into a pattern of black and white squares before you ever look at it. Each square represents a binary value. Black means one. White means zero. Enough ones and zeros in the right order and you can encode almost any information.

When a phone scans the code, the camera captures an image of the pattern. Software running on the phone locates the code in the frame, straightens out any angle or distortion, reads the arrangement of squares in a precise sequence, and converts it back into the original data.

The whole process happens in under a second. From a visual pattern your brain sees as a random arrangement of squares, the phone extracts an exact string of information and acts on it. The code is not random at all, every square has a specific meaning in a specific place.

The first thing the scanning software looks for is not the data, it is orientation. A QR code has three large squares positioned in three of its four corners. These are called finder patterns. Their job is to tell the software immediately where the code is, how big it is, and which way is up. No matter how you angle the phone, or if the code is slightly skewed, the finder patterns let the software correct for it.

On an iPhone, the camera app has had built-in QR scanning since iOS 11. Point the camera at a code and a notification banner appears at the top of the screen, no separate app needed. On Android, the default camera app on most modern devices handles QR codes natively as well, with Google Lens available as a fallback for older models.

Once the finder patterns are located, the rest of the code is a structured grid of data modules. The software reads them in a fixed pattern, not left to right like a book, but in a specific zigzag sequence defined by the QR standard, and converts the binary values back into usable data.

A QR code is a two-dimensional barcode. Where an ordinary barcode stores information in one direction, in the width and spacing of vertical lines, a QR code stores information in two directions: both across and down. This is why QR codes can hold so much more data than traditional barcodes.

The squares in a QR code are called modules. A small QR code might have a 21x21 grid of modules. A large one can have up to a 177x177 grid. Not all modules carry data directly, some form the finder patterns, some indicate the code's format and version, and some carry error correction information.

Error correction is one of the genuinely clever things about QR codes. Part of the data in every code is deliberately redundant, it exists so the code can still be decoded even if up to 30 percent of it is obscured, damaged, or decorated. This is why you can put a logo in the middle of a QR code and it still scans, and why a code on a slightly crumpled receipt usually still works.

A payment QR code is really just a carrier for a payment identifier, a string of data that tells a payment app who is being paid, for how much, and through which system. The QR code itself does not contain money or bank account details in any dangerous sense. It contains a reference that the payment app uses to initiate a transaction through its own secure channel.

In practice there are two models. In one, the merchant displays a QR code and the customer scans it, the code encodes the merchant's payment address, and the app uses that to send funds. In the other, the customer shows a QR code generated by their own banking app and the merchant scans it to pull the payment.

Either way, the QR code is just the handshake that connects two parties. The actual financial transaction happens through the payment network, which has its own security layers entirely separate from the QR code itself.

In every one of these cases, the QR code is doing the same fundamental thing: storing a piece of information that would otherwise require the person to type something out. The code is just a shortcut with better accuracy.

A business card QR code might encode a vCard, a standard format for contact information, so that scanning it immediately adds the person's name, number, email, and company to your contacts without any typing. Or it might simply be a link to a website or LinkedIn profile.

An Instagram QR code encodes a link to a specific profile. Scanning it opens the profile directly in the Instagram app or a browser, bypassing the search step entirely. Wedding invitation QR codes usually link to a wedding website, directions, RSVPs, accommodation suggestions, all in one scan. The code is elegant precisely because it contains nothing except the link, and the link does all the work.

Yes. The scanner only needs a clear image of the pattern. It does not care whether that image is a live camera feed, a screenshot saved to your photos, a picture printed on paper, or a code displayed on someone else's screen.

This is why screenshotting a ticket QR code works at concert venues, transit gates, and airports. The code on screen is the code in the screenshot is the code printed on paper, they are all the same pattern. As long as the image is reasonably sharp and not heavily compressed, it scans just fine.

The important thing is contrast and resolution. A QR code screenshot that is very small, very blurry, or has had its colours altered might fail to scan. But a clean full-size screenshot at normal screen resolution will almost always work as reliably as the original.

The QR code itself has no connection to the internet. It is just a pattern. Decoding what is stored in the pattern requires only the camera and the scanning software, no internet needed for that part.

What happens after decoding depends entirely on what the code contains. If the code encodes a web address, opening that address requires an internet connection. If it encodes a Wi-Fi password, your phone can join the network without any internet. If it encodes plain text or contact details, no connection is needed at all.

So the answer is: QR codes do not need internet to be scanned and decoded. Whether the action that follows requires internet depends on what the code was pointing to.

Yes, with one important condition: the scanner needs to be able to distinguish the dark modules from the light ones. The standard is black on white because that gives maximum contrast, but the actual requirement is just sufficient contrast, not a specific colour.

Dark blue on white works. Dark green on cream works. Coloured QR codes appear on branded packaging, marketing materials, and creative projects all the time. The scanner looks for the difference between lighter and darker areas, not for the specific colours black and white.

White QR codes, light modules on a dark background, also work, as long as the contrast is maintained. What fails is low-contrast combinations: grey modules on a slightly lighter grey background, or pastel colours against white. If the scanner struggles to tell the modules apart, it cannot decode the grid.

Printed QR codes are the most common form in existence, restaurant menus, product packaging, posters, receipts. Any flat print with adequate resolution and contrast scans reliably. The surface does not matter: paper, card, fabric, ceramic, wood.

Tattoo QR codes are technically possible and have been documented working. The challenges are that skin is not perfectly flat, tattoos can fade or distort over time, and skin tone affects contrast. A fresh, high-contrast tattoo QR code on light skin with a skilled tattooist following exact specifications can scan. Whether it still works years later as the ink ages is less predictable.

3D printed QR codes, where the modules are raised or recessed rather than printed in ink, can scan if the depth of the relief creates enough contrast for the scanner to distinguish dark from light under normal lighting conditions. Some work beautifully. Others, depending on the material, surface finish, and lighting angle, confuse the scanner entirely. The error correction built into every QR code gives these unusual versions a fighting chance even when the result is imperfect.

Those three large squares in the corners are not decoration and are not part of the data. They are finder patterns, navigation markers specifically designed to be instantly and unambiguously located by a scanner, no matter the angle, size, or orientation of the code.

Each finder pattern has a distinctive structure: a solid dark square, surrounded by a light border, surrounded by another dark border. This specific pattern of concentric rings does not appear anywhere else in a QR code by design. The scanner's first job, before reading any data at all, is to find all three of them.

Once the scanner finds all three corners, it knows the exact position, size, rotation, and perspective distortion of the code. It can then correct for any angle and lay a precise grid over the data modules. The fourth corner has a smaller alignment pattern for the same reason in larger, more complex codes. These markers are the reason you can point a phone at a QR code from almost any angle, and the scan still works within a second.