How to convert a PDF to PDF/A
Updated
Contents
To convert a PDF to PDF/A, open PDF to PDF/A, add your PDF, press Convert to PDF/A, read any notes, and download the copy. PDF/A is the archiving form of PDF, meant for documents that must still open and look the same years from now, and every job here targets PDF/A-2b.
The copy is rewritten by Ghostscript, the open-source PDF engine, with its PDF/A-2 setting switched on. Conformance level b (basic) is about how the pages look; the tagged structure that level a requires is not added. Nothing checks the result against the PDF/A specification afterwards, so treat it as a best-effort conversion, not a certificate.
PDF to PDF/A runs on a secure server, because that is where Ghostscript is installed. Your PDF, up to 100 MB, travels over an encrypted connection and is deleted once the job is over; the download link stops working after 60 minutes. No account is needed.
Step by step
- 1Open the PDF to PDF/A tool. The badge at the top reads Secure server when conversion is available; otherwise Convert to PDF/A stays off and nothing is sent.
- 2Drop one PDF on the area that says “Add a PDF, or browse”, or click it to choose the file. Remove, beside its name, swaps it for another.
- 3Press Convert to PDF/A. Your PDF uploads over an encrypted connection to our secure server, and the button follows the job while Ghostscript rewrites it.
- 4If Ghostscript reported anything, “What PDF/A could not keep” lists it above “Your PDF/A file is ready.” Read it before you replace your original.
- 5Press Download to save the copy: report.pdf comes back as report-pdfa.pdf. The link stops working after 60 minutes, so save it before you close the tab.
- 6If an archive or portal asked for PDF/A, check the copy with its own checker before you submit it.
What the conversion drops, and how you find out
Some of what a PDF can carry has no place in PDF/A. Ghostscript is told to drop it and carry on rather than abandon the job, and what it reports while doing so is shown above your download in its own words, with any line that names a file path held back. The line you are most likely to meet ends in “discarding DOCINFO”: DOCINFO is the older place a PDF keeps its title, author and producer, and the notice does not say which fields went, so open the copy's document properties if they matter.
PDF/A also requires every font to be embedded. Where your original only names a font without carrying it, Ghostscript uses one from its own set; on our secure server, NimbusSans was seen standing in for Helvetica. The notes do not report that swap, so compare the copy with the original page by page before you rely on it.
What is checked, and what is not
Before the copy is returned, our server checks that it is not empty, that it has a PDF header and end-of-file marker, and that it declares PDF/A-2b: PDF/A identification part 2 and conformance B in its metadata, plus an embedded colour profile. If either declaration is missing, no file is returned.
A declaration is not proof of conformance. If a university repository, court or records system needs PDF/A, run its own checker, or an independent validator such as veraPDF, on the copy. There is no level picker either: a portal that names PDF/A-1b or PDF/A-3b needs other software.
Scans, passwords and damaged files
Converting a scan to PDF/A does not make its text searchable: a scanned page stays a picture of words. Run it through OCR PDF first, which adds a searchable text layer on our secure server, then convert that file and search the PDF/A copy to confirm the words survived.
The page has no password field, so Ghostscript is never given one. If your file asks for a password when it opens, Unlock PDF removes it once you enter that password; then convert the unlocked copy. The PDF/A copy is written without encryption, which PDF/A does not allow in any case.
Our server first looks for a PDF header at the start of your file and an end-of-file marker near its end. A file missing either, such as a download that stopped partway, is refused as one that could not be read as a PDF, and nothing is produced.
Where your PDF goes
Your PDF travels over an encrypted connection to our secure server, where Ghostscript works in a temporary folder that a cleanup step deletes whether or not the conversion worked. The stored upload is deleted once the job ends, or held for a retry if the job hit a temporary error. The PDF/A copy's download link stops working after 60 minutes, and a daily storage rule erases the stored copy after that.
Common questions
Will the converted file pass a PDF/A validator?
It targets PDF/A-2b, but pdf.xyz runs no PDF/A validator. Our server checks that the copy declares the standard and carries a colour profile, and returns no file if it does not; a declaration is not proof, so check the copy with the validator your archive uses.
Can I convert to a different PDF/A level?
No. There is no level picker, and every conversion targets PDF/A-2b. If an archive asks for another level, such as PDF/A-1b or PDF/A-3b, you need software that offers it.
What does “discarding DOCINFO” mean?
Ghostscript found text in the PDF's older information dictionary, where the title, author and producer live, that it could not carry into PDF/A's metadata, and dropped that dictionary. It does not list which fields went, so check the copy's document properties if the title or author matters.
Does PDF/A make a scanned PDF searchable?
No. A scanned page stays a picture of text. Run the scan through OCR PDF first, then convert the result and search the PDF/A copy to check the words came through.
Is my PDF uploaded?
Yes. Ghostscript runs on our secure server, so the PDF travels there over an encrypted connection and is deleted once the job ends. The PDF/A copy's download link stops working after 60 minutes, and a daily storage rule erases the stored copy.