PDF to PDF/A

Beta

Convert a PDF to the PDF/A archival standard (experimental).

or drop PDF here

Runs in your browser — your file never leaves your device.

Convert a PDF towards PDF/A, the ISO standard built for long-term archiving. A normal PDF can depend on things outside itself — a font installed on the reader's machine, a colour profile from wherever it was made, JavaScript, links to files that will not exist in twenty years. PDF/A requires the document to be self-contained so it renders the same decades later. This tool writes the archival structure in your browser: an output intent with a real colour profile, XMP metadata declaring the conformance level, a permanent document ID, and no encryption, scripting or file actions. Then it tells you the one thing it cannot fix, in plain words.

How to use PDF to PDF/A

Click "Select PDF file" and choose the PDF to archive.
Click "Convert to PDF/A". The structural rewrite happens in your browser, so there is no upload and no queue.
Read the font note if one appears. It names any font the original only referenced rather than embedded — because a file that depends on a typeface the reader does not have is not truly self-contained.
Download the converted file and keep it beside the original, not instead of it.

Who it is for

PDF to PDF/A is for documents that have to still be readable in ten or twenty years: records that a registry or court will accept, a thesis deposited with a university library, a contract archive, a publication deposited with a national library. It is also used by anyone told by a body that a submission "must be PDF/A" and who needs to get as close as possible without a copy of Acrobat.

How it works

The document-level requirements are written exactly: an output intent naming a colour space with a real ICC profile embedded, XMP metadata declaring the part and conformance level, a permanent document identifier, and the removal of encryption, JavaScript and launch actions.
Fonts are inspected rather than silently ignored. Any font the PDF only referenced rather than embedded is named, because a file depending on a typeface the reader does not have is not self-contained.
Everything runs in the browser, which is what removed the old silent failure where a server-side converter returned blank pages.

What it cannot do

A font that is only referenced cannot be embedded here — that would need the typeface itself and a layout engine. The file is improved and still opens everywhere, but a strict validator will reject it and the tool says so.
It cannot promise certification. Conformance is a validator's judgement, and the honest description is "PDF/A-structured".
Encryption is removed as part of the profile, so an encrypted input has to be unlocked first.
Keep the original alongside the archival copy; the conversion is a structural rewrite, not a substitute for the source.

Frequently asked questions

Is the output certified PDF/A?

It is PDF/A-structured, and that is the honest description. The tool writes the document-level requirements exactly — colour output intent, XMP metadata, permanent document ID, and the removal of encryption, JavaScript and launch actions. What it cannot do is embed a font the original only referenced, because that typeface is simply not present in the file to copy. Whether the result is strictly conforming depends on the fonts, and the tool tells you which side of that line your file fell on.

So which fonts does it complain about?

Ones the PDF measured by reference rather than carrying inside itself — often a licensed corporate face, or text exported without font embedding. Most files produced by Word, InDesign, LaTeX and the major PDF exporters already embed everything, and those convert cleanly with no note at all. When a font is missing the note names it, so you know to check that page in another viewer rather than assume it will look the same everywhere.

What do I do if a font is not embedded?

Two options. Accept it, if the document only needs to be archived rather than strictly validated — the structural improvements are still real. Or replace that text with one of the nineteen faces this site can embed (three built in, sixteen downloaded on demand), by editing the page before converting, and then run the conversion again against a file that has no missing fonts.

Why did the old version of this tool return blank pages?

It shelled out to Ghostscript with an archival profile definition, and on this hosting that combination failed by producing a PDF with no pages rather than raising an error — so the download was blank white sheets. The conversion now runs in the browser and needs nothing installed on the server, which removes that whole class of failure.

Is my file uploaded?

No. The document is read and rewritten in your browser, so it never reaches a server. That also means the only size limit is your own machine's memory.