QR Code Generator

How to use the QR Code Generator

1. Type or paste the exact destination. For a link, paste the full URL including https:// — the code stores the characters you give it, so a missing scheme or a typo is stored too, faithfully, forever.
2. Pick an error-correction level. Start with M for screens and ordinary print. Choose Q or H when the code will live somewhere rough — a poster outdoors, packaging, a sticker that will get scratched — and L only when the content is long and the code will be shown large and clean.
3. Watch the character and byte count. The counter under the form shows characters, UTF-8 bytes, and how many bytes the chosen level can still hold. If the input is too long for the level, the tool tells you the exact limit and which stronger-capacity level would fit instead.
4. Choose a size and download. Take the PNG for documents, slides and websites, at 512 px for ordinary use or 1024 px when it will be printed large. Take the SVG for print work: it is built from the same module matrix and stays sharp at any size.
5. Scan-test the file you downloaded. Open it on another device or print one copy at the real size and scan it in ordinary light. A code that scans on your desk may fail on a wall, in glare, or at a distance — testing the final artefact is the step that catches that.

Worked example

The sample loaded by the button below is the address of this very tool page. It is short, plain ASCII, and it shows how the counting works before you trust the tool with anything longer or more complicated. Select Load sample and the workspace fills with that URL at level M and 512 px, and the counter reports the figures in the box — every one of them calculated by the generator’s own code, not typed in by hand:

Text: https://atoztools.pro/tools/qr-generator/
Characters: 41
UTF-8 bytes: 41
Level M capacity: 2331 bytes — fits, 2290 bytes spare
Level H capacity: 1273 bytes — fits, 1232 bytes spare
Download filename (PNG): qr-code-atoztools-pro-tools-qr-generator.png

Because the text is plain ASCII, characters and bytes are the same number: 41 of each. That stops being true the moment the text uses accented or non-Latin characters. The word café is 4 characters but 5 bytes in UTF-8, because the é takes two bytes — and QR capacity is counted in bytes, in the byte mode this tool plans against. A long message in Arabic, Urdu or Chinese can therefore reach a level’s byte limit while the character count still looks modest, which is exactly why the counter shows both numbers. Note also what the spare capacity means in practice: this 41-byte URL fits even at level H, the densest and most damage-tolerant setting, so for a short link there is no capacity reason to accept the weaker correction of L or M — the only trade-off that remains is how dense the printed code looks.

What a QR code stores, and how error correction trades against density

A QR code stores characters — nothing more. The black and white squares (modules) are grouped into a grid whose size grows in steps called versions, from version 1 at 21 by 21 modules up to version 40 at 177 by 177. There is no server behind a QR code, no account, and no way for the code itself to know it was scanned. When a phone reads one, it simply decodes the stored text and then decides what to do with it: open it as a link if it looks like a URL, show it as text if it does not. Anything beyond that — a redirect, a scan counter, a changed destination — has to live on a server the code points at, which is why this tool is explicit about what it does: the code it generates encodes exactly the text you typed, character for character, with no short link or tracking service inserted in between. If the destination ever needs to change, the honest options are to reprint the code or to point it at a landing page you control from the start.

Error correction is the code’s insurance against the real world. Alongside your data, the encoder writes redundant information calculated with Reed–Solomon codes, so a reader can reconstruct the message even when part of the grid is dirty, torn, faded or covered. The four levels — L, M, Q and H — write increasing amounts of that redundancy. The trade is unavoidable and worth understanding: redundancy takes room, so at any given version a higher level holds less of your data, and holding the same data at a higher level pushes the encoder to a bigger, denser version. The version 40 byte-mode ceilings used by this tool are 2,953 bytes at L, 2,331 at M, 1,663 at Q and 1,273 at H. Those are ceilings, not targets: a code anywhere near them is a tight field of tiny modules that many cameras, especially older ones in poor light, will struggle to read.

That is why the practical advice runs the other way from the numbers. Keep the encoded text as short as the job allows — a clean landing URL rather than a paragraph of parameters — take the strongest correction level that still fits comfortably, print or display the code large enough that each module is clearly separated, and keep the quiet zone, the blank margin around the grid, intact when you crop or place the file. Contrast matters too: dark modules on a light background is the combination readers are built for, and inverting it or setting the code over a photograph is a reliable way to make a technically valid code that nobody can scan.

Common errors, and the limits of this tool

The most common error is encoding the wrong text and discovering it after printing: a URL with a typo, an old domain, a link missing its https://, or a page address that redirects somewhere the reader did not intend. The second is overloading the code — pasting a long paragraph, a whole vCard or a URL stacked with tracking parameters — until the grid is so dense it only scans at arm’s length in perfect light. The third is a layout error rather than an encoding one: codes printed too small for the scanning distance, cropped into their quiet zone, stretched out of square, or placed on curved or glossy surfaces that reflect the camera’s light back at it.

The limits are these. Capacity is a hard ceiling per level — 2,953 bytes at L, 2,331 at M, 1,663 at Q and 1,273 at H, counted as UTF-8 bytes — and no setting in this tool can extend it; longer content must be shortened or hosted at a short address the code points to. At level H in particular, even a few hundred bytes produce a dense code, so for print, choose size by scanning distance: the further away the reader stands, the physically larger the code needs to be, and a code that will be scanned from across a counter should be tested from across a counter. The tool draws the code and builds the PNG and SVG downloads in your browser from the text you enter; nothing is uploaded. It cannot verify that the page a URL points to exists, that it is safe, or that it will still be there next year — check the destination yourself, and remember that a QR code hides its destination from the person scanning it, which is a reason to use codes honestly and a reason readers are right to be cautious about strange ones.

Frequently asked questions