PDF/A Converter

PDF/A conversion is refused here — this page says why, and what part of an archiving job can still be done.

Secure server processingPDFWorkingNo account needed

What this tool does

PDF/A Converter is registered, and it does not convert. Selecting it returns a refusal with the reason attached, because ISO 19005 conformance is not a label that can be stapled to a document: it requires an embedded ICC output intent, every font subset carried inside the file, no encryption, and no reference to anything outside itself. Producing that combination needs a colour-profile engine this server does not run.

The engines that do the work — Ghostscript and MuPDF — are offered under AGPL or commercial terms this project has not taken on. What it does run is qpdf and poppler on the server, for the few PDF jobs a browser cannot manage, and pdf-lib and pdf.js inside your own tab for most of the rest; not one of those four converts a colour space or embeds an ICC profile. The shortcut would be to add the XMP block declaring PDF/A-2b and hand the file back unchanged. A validator would reject it in the first second of checking. That is the file we would rather not make.

The rest is still worth knowing, because conversion is the last step of archiving rather than the whole of it. A scan with no text layer can satisfy PDF/A-2b and stay unsearchable, and asking for 2u instead will not change that: its Unicode rule governs the text a file already carries, not text that was never captured. A document with a damaged cross-reference table fails long before colour is considered. Both of those are fixable here, with PDF OCR and Repair PDF, and fixing them first turns the eventual conversion into a formality instead of a sequence of rejections.

How to use it

  1. Open the tool and read what it reports. The workspace asks the service which operations can run before it draws anything, so what stands where a dropzone would be is the refusal itself, carrying the reason the server gives for it.
  2. Find out which conformance level you owe. An archive, a court or a funder names one: PDF/A-1b, 2b, 2u or 3b. Delivering the wrong level is refused as firmly as delivering an ordinary PDF.
  3. Fix what is fixable here first. PDF OCR gives a scan a real text layer, Repair PDF rebuilds a file whose structure a converter would choke on, and PDF Metadata shows the title and author a validator will inherit.
  4. Convert elsewhere, then validate. A converter that does not check its own output is a guess. veraPDF is the open-source validator maintained with the PDF Association, and its report is what an archive will accept as evidence.

Processing & privacy

Secure server processing. This operation requires temporary server processing.

Common use cases

A registry or court accepts nothing else

Electronic filing rules often name ISO 19005 and a specific level. This tool cannot get you there, so budget for a converter that ships a validation report with the file.

Depositing research outputs

Data-management plans ask for archival PDF because the repository has to open the file decades from now. Embedded fonts are what make that true, not the file extension.

Digitising a paper file

Scanning is the hard part, and it can be done here: PDF OCR adds the text layer that separates a searchable archive from a stack of pictures with a PDF wrapper.

An invoice carrying an XML attachment

Factur-X and ZUGFeRD invoices are PDF/A-3 files with machine-readable data embedded inside them. That level exists to permit the attachment: PDF/A-2 will embed only another conforming PDF, and PDF/A-1 no file at all.

Questions

Can I convert a file to PDF/A here?
No. The operation is registered so the site can refuse it precisely, naming the missing engine, rather than accepting the job and returning an ordinary PDF that has been quietly relabelled.
Why not just write the PDF/A metadata and be done?
Because conformance is structural, not declarative. The XMP block only states a level. The document still has to embed its fonts, carry an ICC output intent, avoid encryption, stay inside the features its level permits, and reference nothing external. Declaring without doing produces a file that fails on the first check.
Which conformance level should I ask for?
PDF/A-1b has the widest acceptance and the tightest rules — no transparency, no optional content layers. PDF/A-2b relaxes both, 2u adds the requirement that all text maps to Unicode, and 3b is the only level that allows arbitrary embedded attachments.
Is my file uploaded when I open this page?
No, because the refusal takes the place of the file picker. One gap is worth naming rather than hiding: should that availability check itself fail, the page falls back to drawing a dropzone, and a document dropped into it is uploaded before the API turns the job down — it is then never processed, and the retention sweep deletes it. Tools here that do run on the service send the document over an encrypted connection and it goes the same way afterwards.
Will this tool start working later?
Only if a colour-profile engine is added to the server, which means accepting the licence terms of Ghostscript or MuPDF. Until that decision is taken and the conversion is genuinely implemented, the answer stays no.
What can be done here to a document headed for an archive?
PDF OCR adds a searchable text layer, Repair PDF rebuilds a damaged file so a converter can read it, PDF Metadata shows what a validator will see, and Compress PDF shrinks the file without discarding the text.

Related tools

  • Merge PDF — Combine several PDFs into one document in the order you choose.
  • Split PDF — Split one PDF into separate documents by page range.
  • Compress PDF — Reduce PDF file size while preserving the quality that matters.
  • PDF to JPG — Export every page of a PDF as a JPG image.
  • JPG to PDF — Turn one or many JPG photos into a single PDF.
  • Protect PDF — Add a password and set permissions on a PDF.