PDF Logic Team · Checked

PDF checks you can inspect

These 6 examples show what happened when we sent synthetic PDFs through the local application’s merge, split, rotate, compression, redaction and PDF/A tools. Download the input and output files to inspect the same pages yourself.

They demonstrate specific behaviors on specific files. They do not establish a success rate for other documents, processing speed or security certification.

What is in the sample files?

Sample A has three labeled pages with different page sizes and one synthetic text form field on page 1. Sample B has one labeled page. The compression input contains a generated color pattern, and the redaction input contains a fictional secret beside public text. None contains personal information.

Each example links its own inputs below. Keep the originals so you can compare changes and try different settings.

Observed in this sample

Merge two PDFs in input order

Settings: Sample A (3 pages), then sample B (1 page)

  • 4 pages in the requested order.
  • The sample field value remains extractable as static text; no interactive fields remain.

Output page labels: SAMPLE A PAGE 1 → SAMPLE A PAGE 2 → SAMPLE A PAGE 3 → SAMPLE B PAGE 1

Observed in this sample

Extract pages 3 and 1 into one PDF

Settings: Page ranges: 3,1

  • 2 pages in one output PDF, ordered as requested.

Output page labels: SAMPLE A PAGE 3 → SAMPLE A PAGE 1

Observed in this sample

Rotate a three-page PDF by 90 degrees

Settings: Angle: 90 degrees on all pages

  • All three page rotation values are 90 degrees.
  • The original page dimensions and text labels remain.

Output page labels: SAMPLE A PAGE 1 → SAMPLE A PAGE 2 → SAMPLE A PAGE 3

Observed in this sample

Compress a synthetic image and compare the result

Settings: Recommended preset; no Desired Size target

  • Input: 550,146 bytes. Output: 273,956 bytes.
  • Reported outcome: compressed; method: adaptive.
  • The output has one page and its heading remains extractable. Compare the pattern at high zoom: these checks do not establish identical pixels or lossless compression.

Output page labels: SAMPLE IMAGE PAGE 1

Observed in this sample

Remove a fictional secret while retaining public text

Settings: Page 1; left 8%, top 22%, width 84%, height 7%; black fill

  • The fictional secret is absent from extracted output text.
  • The public heading and the sentence below the selected region remain extractable.
  • This checks one visible text region; it is not a guarantee that an arbitrary sensitive document has been fully sanitized.

Output page labels: SAMPLE REDACTION PUBLIC

Observed in this sample

Convert one sample to validated PDF/A-2b

Settings: PDF/A-2b conversion with the configured RGB output profile

  • The application returned the file only after veraPDF PDF/A-2b validation passed.
  • The output retains one page and its sample label.
  • Conformance does not establish accessibility, legal acceptance or identical appearance.

Output page labels: SAMPLE B PAGE 1

How to repeat the checks

  1. Choose an example and download its input files. Keep those copies unchanged.
  2. Open the linked tool and choose the settings shown. For redaction, the percentages describe the rectangle relative to the visible page, measured from its top-left corner.
  3. Download your output and compare its page labels, order and appearance in a PDF viewer. Check whether the sample form value is static after merging, compare fine image detail after compression, and search the redacted file for the fictional secret.
  4. For PDF/A, check the requested conformance level as well as the visible content. A successful conversion is not evidence that a different archival format or accessibility requirement is met.

The automated checks inspected PDF structure and extracted page text. The PDF/A conversion also required veraPDF validation. They did not compare every rendered pixel. Different application versions can produce different files; the observations and downloadable outputs here belong to the dated run above.

What these examples do not cover

These inputs do not represent encrypted, damaged, scanned or large documents. The checks do not establish bookmark preservation, accessibility, general redaction safety or output quality for other documents. PDF/A conformance applies only to the identified archive output. The split example does not check interactive form behavior. Inspect your own results before relying on them.

See our editorial explanation for how capability claims and corrections are handled.

Method and file fingerprints

Local HTTP application endpoints; check runner Node v24.2.0; synthetic PDF inputs. Actual HTTP POST requests; PDF structure read with pdf-lib and text extracted with unpdf. The PDF/A endpoint requires independent veraPDF validation before returning a file. Compression measurements come from response bytes and checked headers. Output hashes identify this run, not universal byte-for-byte reproducibility.

SHA-256 fingerprints let you confirm that downloaded files match this published run.

sample-a.pdf · 2,547 bytes
04c7ab483328931e598a922ec0c724e2e3c25249ccd6a15321ce44769bb0de4b
sample-b.pdf · 993 bytes
fdd9f9c5c9aa0bdd39a411c47bd8378362604e691c0d7e23484078c75f532770
sample-image.pdf · 550,146 bytes
674ffae63a7e852304b105a576981a5ebf437a838e0a5bb3eaf0bfc849c9e641
sample-redaction.pdf · 987 bytes
247fdf97fa494593e7cf77807193c11e57a067024ca0fa1fe645ea0476094ccd
merge-output.pdf · 2,525 bytes
ed45b301680abe3dbc3f95aed596523fe56a8cef0e408d965770814038b04e27
split-output.pdf · 1,822 bytes
b8d73d5913d7e4045ce522b126685a8d922db64c3f6dbe8416cb5bc5e00f1d13
rotate-output.pdf · 2,561 bytes
cbfd42fd963d303a5462aea4aafd9488d34301c06b3b304be373f8a23e2aae51
compress-output.pdf · 273,956 bytes
8b59ce5eaf57c473668bb1747e861de56a59ea676fba2bda4b5cfa2cfa48930a
redact-output.pdf · 920 bytes
bb1c5d45ebe718043c527ddac8353013ef0d982a80a19f467f9349c1d560e775
pdfa-output.pdf · 7,927 bytes
6f06b63d7bcd53a702dac2041020219243ef0083eca16244daf765bef778e615
Download the complete results and fingerprints