Skip to content
QRMakery

Round QR Code Generator

Rounded modules or separate dots, with the three corner squares left square, because those are what the scanner finds the code with.

The page opens on rounded modules at High error correction. That pairing is not decoration: rounding spends margin, and High is what puts it back.

Add your link or content

What the code should contain

Only http and https links are allowed. Missing schemes become https.

Customize your QR code

Finder squares stay square whichever you pick. They are what a scanner locates the code with.

The blank margin scanners need. 4 is the standard minimum.

Download your QR code

✓ Scan-tested: this code reads back correctly.

Contrast 21:1, Strong

Version 3 · 37 modules · print at least 14.8 mm wide

Rounding Is Not a Style Change

A module is a square of ink, and a scanner decides whether it is dark by sampling inside it. Round the corners and you remove ink from that square. Separate the modules into dots and you remove more, and you open gaps between neighbours that the scanner can sample into by mistake.

That makes shape part of the same budget as contrast, logo coverage and print size. Here is each shape at each error-correction level, encoded and read back at build time.

  • Square

    Full module area. Nothing is given up.

  • Rounded

    Corners clipped, horizontal runs stay joined.

  • Dots

    Every module separated. The most inked area lost.

ShapeLevel LLevel MLevel QLevel H
Squarereads backreads backreads backreads back
Roundedreads backreads backreads backreads back
Dotsreads backreads backreads backreads back
The same link encoded twelve ways and read back with the decoder the download button uses. A cell that says it reads back was proved, at this payload length, on a clean render. Print, ink spread and a phone camera all take margin off that, which is why the shape you pick should not be the thing spending your last few percent.

Why the Three Corners Stay Square

The big squares in three corners are the finder patterns, and they are how a scanner locates the symbol and works out its orientation and scale. A scanner sweeps a line across the image looking for a specific run of dark and light in the proportion 1:1:3:1:1. That measurement depends on where the edges are.

Round those edges and you move the thing being measured, which degrades detection before it degrades decoding, so the code gets harder to find rather than obviously broken. This tool keeps the finder patterns square in every shape, which is why a dotted code here still has three crisp corners. That is not a design choice you can override, and generators that let you round the corners too are selling you a code that is harder to find.

Getting Away With It

Take the SVG rather than the PNG for a shaped code. Rounded corners and circles are exactly the detail a raster export softens first, and softened edges are what a scanner misreads. If the design also needs no background box, the transparent tool tests each surface separately, which matters more once the modules are already giving up area. For the full identity, the branded code tool combines shape, colour and a logo with every check applied at once.

  • Prefer rounded over dots. Horizontally adjacent modules stay merged into one shape, so a run of five reads as a bar with soft ends rather than five islands.
  • Raise the error correction when you change the shape. High if you are using dots, and do not spend the same budget twice.
  • Do not stack effects. Dots plus a logo plus a light brand colour is three withdrawals from one account.
  • Print larger than the minimum shown under the preview. Shaped modules have less ink to lose before they stop registering.
  • Proof it on the real printer. Ink spread closes small gaps, which sometimes rescues a dotted code and sometimes smears it, and that depends on the paper.

What This Does Not Do

  • No custom finder-pattern styles. They stay square, for the reason above.
  • No per-module colours or gradients. Solid colour only, so the contrast measurement means something.
  • The decode test runs on a clean render at generous resolution. It is the easiest test a code will ever face, so treat a pass as necessary rather than sufficient.
  • It cannot make a shaped code as reliable as a square one. The square is the shape the format was designed around.

Rounded QR Code Questions

Do rounded QR codes still scan?

Yes, when the rest of the code has margin to spare. Rounding removes inked area from every data module, so it spends the same budget that low contrast and small print sizes spend. The grid on this page shows each shape decoded at each error-correction level.

Why are the three corner squares not rounded?

Because they are what the scanner finds the code with. The finder patterns have a fixed 1:1:3:1:1 ratio measured across a scan line, and rounding moves the edges that measurement depends on. This tool keeps them square in every shape, deliberately.

Are dots or rounded modules safer?

Rounded. Horizontally adjacent modules stay merged into one shape, so a run of five reads as a bar with soft ends. Dots separate every module, which loses the most inked area and is the shape most likely to fail in print.

What error correction should I use with dots?

High. Dots spend margin, so give the code the largest recovery budget available. If you also want a logo or a brand colour, reconsider dots: those spend from the same budget, and three at once is where codes stop working.

Will a rounded code print as well as a square one?

Print at the width shown under the preview or larger, and test the actual print. Ink spread on paper closes small gaps, which sometimes helps a dotted code and sometimes blurs it. This is the one design choice worth proofing on the real printer.

Colour, a logo in the middle, artwork behind it, each one measured against what still scans.

All tools

Guides worth reading

All guides
  • Comparisons

    Static vs Dynamic QR Codes

    What the redirect actually buys you, measured on real encoded codes, and which payloads are worse off for it.

    16 min read