Skip to content
Label2A4
All practical guides

Transparency and quality

How Label2A4 detects and verifies a shipping label

This method describes the current product, the evidence used to approve a change and the situations where a human check is still required. It is not a claim of universal carrier certification.

By

Label2A4 creator · 8 min

Sources last reviewed :

1. Test scope

A test starts with an untouched carrier PDF. The set deliberately covers single-label pages, several labels on one page, rotated content, large blank areas, embedded images, advertising or instruction pages and documents containing several code-like patterns. Personal shipping data is never published as an example. Synthetic or masked fixtures are used in automated tests.

2. Detection sequence

The browser opens the file with PDF.js. Label2A4 first uses embedded or visually detectable codes as strong location clues, then analyses non-white content and known carrier layouts. It can split a page containing several labels, normalise the crop, rotate it and fall back to a larger content or full-page area when confidence is insufficient. A carrier profile and manual crop remain available.

3. Acceptance criteria

A result is accepted only when the preview retains the full main barcode or QR code, its clear margin, the human-readable tracking reference and the useful destination block. It must not select a blank page, advert or isolated partner reference as if it were the parcel label. Label order and output page count are also checked. A physical print must also be checked on the sender’s printer after a scaling or layout change.

4. Automated and manual tests

Automated tests protect geometry, rotations, duplicate handling, page composition and known parsing cases. Build and lint checks catch route, type and rendering problems. A complementary manual check should compare the original, in-app preview, downloaded PDF and an actual printed A4 sheet. No single test proves acceptance by every scanner, so carrier requirements and a sender's final visual check remain necessary.

5. Privacy during processing

The normal PDF preparation path runs in the browser. The source file does not need to be stored on Label2A4 servers to create the output. Authentication, quota, purchases and usage measurements use separate service data described in the privacy policy. When support needs a failing file, the user chooses to send it; it is not silently collected from the tool.

6. Paper-savings calculation

The calculator compares one A4 sheet per label with up to four labels per generated sheet. Sheets used equal the ceiling of labels divided by four; sheets avoided equal labels minus generated sheets. For 40 labels, this gives 10 generated sheets and 30 avoided sheets, or 75% fewer sheets in that full batch. This is a material-use estimate, not a certified carbon claim.

7. Known limits and reporting

A new carrier redesign, a low-resolution scan, an encrypted file, a damaged download or an unusual multi-document page can defeat automatic detection. Manual crop is the safe fallback. Reproducible cases are prioritised when support receives the untouched PDF, carrier, expected number of labels and the observed step. The method and guides are updated when the product changes or an official rule is revised.

8. Editorial policy

Every practical article names its author and review date. First-hand product behaviour is separated from carrier rules, which link to official documentation. Similar carrier pages receive carrier-specific checks instead of changing only brand names. Limitations are stated plainly and older content is corrected rather than kept solely to attract search traffic.

What the software checks actually cover

The suite uses synthetic documents and local account/advertising responses. The following scenarios are reproducible software checks, not physical carrier acceptance tests.

Chronopost small print
The fixture places several lines below the main code and a receipt alongside it. The exported A4 is rendered and the test counts visible ink in each expected line; text extraction alone is not treated as proof that a clipped line is visible.
Manual crop and rotation
An explicit manual crop must not be automatically enlarged. Rotated pages and separate label frames retain their selected areas and expected counts.
Workspace continuity
Import a synthetic PDF, choose manual mode, visit a guide and resume the batch. The test checks the retained file, setting and preview. Advertising is suppressed while a workspace remains mounted.
Consent and empty pages
Local stubs test missing consent, partial purposes, refusal, paid access, errors and unknown guide paths. These simulations do not attest that a real Google message has been published.

Still required on real equipment: a physical print, a ruler measurement where relevant and the carrier’s operational checks. No universal scanner success rate is claimed.

A concrete acceptance checklist

  • No useful edge or quiet zone is clipped.
  • Recipient and sender blocks are not confused in the preview.
  • A tracking fragment containing another brand does not override the visible carrier layout.
  • Rotated and multi-label pages keep the expected count and order.
  • Blank or advertising-only pages do not become parcel labels.
  • The downloaded PDF matches the approved preview and prints as one A4 page per sheet.

Sources and further reading

Carrier rules remain authoritative. Interface screenshots do not replace checking the current requirements of your chosen service.