How to judge a QR code generator

The eight things that actually decide it

One: is the code static or does it route through a service. Two: what happens to the free tier over time. Three: which export formats exist, and whether SVG is one of them. Four: can you see the payload before you print. Five: can you set the error correction level. Six: where your data goes, meaning whether the text you type is stored. Seven: how the destination is expressed, as a URL or through a shortcode. Eight: can you test the result before the print run.

A generator that scores well on all eight is good for almost any job. One that scores badly on the first and third is a service with a generator attached, which is a different product.

• Static payload you can read in a text editor
• SVG export, not only PNG
• Error correction level visible and changeable
• No account needed for a plain code

Test one: look inside the code

Generate a code for a URL you control, download it as SVG, and open the file in a text editor. If the payload contains your URL, the code is static. If it contains a short address belonging to the generator, it is a redirect, and everything about how long it lasts depends on that company.

This single test separates most of the market into two groups faster than any feature page.

Test two: read the export, not the preview

Previews are rendered at screen resolution and cropped kindly. The export is what reaches the printer. Check the pixel dimensions of a PNG, or better, whether an SVG exists, and check whether anything has been added to the image: a frame, a logo, a caption.

Then check the quiet zone. Four modules of clear space belong around the pattern, and a generator that adds a decorative frame inside that margin has made the code worse in exchange for a nicer thumbnail.

Test three: try to break it before printing

Print one copy at the size you plan to use, put it where it will actually live, and scan it with two phones in that light. Then scan it with the phone at an angle, because people scan at angles.

This costs a sheet of paper and catches the failures that no comparison table mentions: too small for the distance, too little contrast on the material, gloss reflecting a window.

Test four: ask what the free tier really is

Free static codes and free trials look identical on the first day. They differ on the day the trial ends, when one keeps working and the other becomes a printed dead end. Read the terms for the word expiry, and check whether the generated file contains a redirect before you commit to a print run.

A generator that offers editing and scan counts for free is selling something other than codes. That is not a criticism, it is a description of the business, and knowing which business you are dealing with is the whole point.

Test five: check the small things that annoy later

Does it handle a URL with an umlaut or a space correctly, encoding it rather than truncating it. Does it strip spaces from a phone number before encoding it in a tel link. Does it warn you when a Wi-Fi password contains a character that breaks the payload. Does it let you choose the error correction level for a code that will carry a logo.

These are the differences between a tool that produces something plausible and one that produces something that works in a hallway with bad light.