Getting a QR code to scan off a printed pack, first time and every time
Getting a QR code to scan off a printed pack, first time and every time
- Open source
- No sign-up
- No watermark
- Free API + MCP
- In the MCP registry
What print mechanically does to a code
Three mechanisms do the damage. Ink gain spreads every mark outward, so dark squares grow into their white neighbours and the quiet, crisp grid you authored becomes a slightly blurred one; the smaller the modules, the more of the code gain eats. Misregistration shifts print units relative to each other, which only matters for a code if the code is not printed as a single solid colour, and is one more reason it should be. Substrate and finish change the surface itself: fibrous corrugate scatters the edge of every mark, gloss laminate turns raking light into a mirror, and flexible film creases in transit exactly where the code ended up. Error correction level H buys back a surprising amount of this, but it buys back damage, not design: it cannot rescue a code that was printed too small or too low-contrast in the first place.
| Process | Typical behaviour with codes | What to ask the printer |
|---|---|---|
| Digital label | Crisp small modules; the safest process for tight panels and variable data | Confirm the code prints as one solid, not a build from multiple toners |
| Flexo | Pressure and plate softness spread marks; the commonest cause of filled-in small modules | Printed proof on the real stock, and a module size the plate room is happy with |
| Offset litho on carton | Fine detail well, but watch reverse-out panels and heavy coverage near the code | A matte varnish or matte block over the code area if the pack is gloss-laminated |
| Direct print on corrugate | Coarse substrate, spread and wick; the least forgiving common surface | Bigger code, thicker modules, solid dark ink; proof on the actual board grade |
| Gravure / film | Excellent detail, but the film creases and the surface glitters under store lights | Matte patch or orientation choice; higher error correction for the crease risk |
Contrast is a property of the panel, not the logo
A scanner thresholds the image into dark and light, so what matters is luminance difference, not brand colour. A mid-teal code on white can pass; the same teal reversed out of a navy panel will not, because the code is now the pale thing and the panel is the dark thing. The dependable convention is a dark code on a light patch, printed as one solid, with a quiet zone of clear substrate around it that carries no artwork, no border, no pattern fill. Where the pack's finish is gloss, ask for a matte varnish or matte block over the code area: glare is contrast loss at the exact moment of scanning. And keep reverse-out codes, white squares on a solid dark panel, off the artwork entirely; a few scanners handle them, most of the field does not, and the failure is found by a customer in a shop.
Size: distance sets it, substrate raises it
The size guide's method holds: roughly a tenth to an eighth of the scanning distance as a starting rule, with phones doing the reading at arm's length for most packs. Then let the substrate argue the floor upward. A cosmetics label can hold a small code; corrugate wants visibly bigger modules for the same reliability; anything on a curved or flexible surface wants margin for the distortion it will meet. Error correction H densifies the grid, so when you raise the level, raise the size with it. The order of operations that works: fix the destination, pick the error level, then size the code to the worst panel it will be printed on, not the best.
Curves, sleeves and the flattest panel rule
A code reads best when the phone sees it flat. On bottles and tubes, place the code on the flattest panel available, sized so it wraps a modest arc rather than running around the vessel, and keep it away from shoulders and bases where curvature turns sharp. Shrink sleeves are the sharp end of this: the sleeve is printed flat and then distorted onto the form, so a code printed naively arrives stretched. Sleeve printers work from preform distortion data and will place and pre-distort the code if asked, which is a conversation for specification, not a surprise for the proof. Pouches and film add crease risk in transit: a code on the crease line meets a fold every time the pack is handled, which is a placement decision before it is a print decision.
The press proof that predicts the shelf
- 1
Proof on the real stock, on the real process
A digital proof of the artwork is not a print proof. Same substrate, same press conditions, same finish.
- 2
Scan dry, then scan wet
Condensation on a chilled pack is a real scanning condition. Know how the code behaves with a wet surface before the customer finds out.
- 3
Scan under shop lighting, then phone torch
Gloss glare and dim ambient are the two common failure lights. The torch also simulates a customer's workaround.
- 4
Three devices minimum
Different camera and decoder combinations disagree at the margin. Pass on the newest phone, a mid-range one, and something older if the audience includes it.
- 5
Sign the proof and file it
The signed press proof is the arbiter when a reprint drifts. Without it, every reprint re-litigates the code.
Put the code into artwork as vector, at final size, once
The quiet zone and the module edges are the first casualties of a code that gets exported as a low-resolution PNG, dropped into a layout, scaled up and re-saved. Hand artwork an SVG sized in millimetres for that panel, generated in the same batch as the rest of the run, and the code arriving at the press is the code you specified. Rescaling is where quiet zones quietly die.