Convert to PDF/A — Free, In Your Browser

PDF/A is the long-term archival flavour of PDF — a strict subset designed so a file you save today will render identically in 50 years, on viewers that don’t exist yet. Courts, government archives, and financial regulators ask for it. Most PDFs aren’t compliant out of the box. Signegy adds the metadata that turns a compliant-by-content PDF into a compliant-by-spec one. Free, no signup, your file never leaves your browser.

What PDF/A Is, In One Paragraph

A normal PDF can reference a font by name (Helvetica, Times) and trust the viewer to substitute. It can rely on system colour profiles, embed encrypted streams, link to external multimedia, and use features that PDF readers might interpret differently. PDF/A removes all of that. Every font must be embedded inside the file. Every colour must resolve through an embedded ICC profile. Encryption is forbidden. External references are forbidden. The result is a self-contained file that any conformant viewer, today or in 2075, will render the same way.

When PDF/A Is Required

PDF/A appears in three buckets of real-world use:

• Legal and government archives. Most national archives accept only PDF/A. The German Federal Archive, the U.S. National Archives, the EU’s Court of Justice document portal — all require PDF/A-1 or PDF/A-2. Court e-filing systems in many U.S. states (PACER for federal civil) accept regular PDF, but state-level systems increasingly require PDF/A.
• Long-term financial records. The German tax authority requires invoices retained for 10 years to be PDF/A. The Swiss tax framework requires PDF/A-3 specifically (so structured invoice XML can ride alongside the rendered version). Many enterprise document-management systems (OpenText, M-Files, Alfresco) auto-convert ingested PDFs to PDF/A on the way in.
• Standards-driven workflows. ISO 19005 (the standard PDF/A is defined by) is referenced by industry-specific specs in pharma (TMF Reference Model), construction (ISO 16739 / IFC), and engineering (LOTAR — long-term archiving and retrieval). If you’re shipping documents into one of those workflows, PDF/A isn’t optional.

If you’re emailing someone a PDF for them to print, sign, or fill — you don’t need PDF/A. If you’re depositing a PDF into a system that has to retrieve it readably 30 years from now, you do.

The Variants: PDF/A-1, A-2, A-3

PDF/A has three versions, each based on a different underlying PDF version, with two conformance levels (a and b):

Variant	Year	Base	Adds	Notes
PDF/A-1b	2005	PDF 1.4	—	Most widely accepted; what we produce.
PDF/A-1a	2005	PDF 1.4	Full Unicode, structure tags	Accessibility (screen readers); needs tagged source PDF.
PDF/A-2b	2011	PDF 1.7	JPEG 2000, transparency, embedded files	Smaller files for image-heavy content.
PDF/A-2a	2011	PDF 1.7	Same as 2b + tags
PDF/A-2u	2011	PDF 1.7	Unicode-mapped text	Halfway between 2b and 2a.
PDF/A-3	2012	PDF 1.7	Arbitrary file attachments	Used by ZUGFeRD / Factur-X invoicing.

The b levels (basic) certify the visual rendering is preserved. The a levels add accessibility — every glyph must map to Unicode, document structure must be tagged. The a levels are much harder to produce because they require the source PDF to already have proper structure tags. Most PDFs in the wild don’t.

We target PDF/A-1b: maximum compatibility, the easiest gate for an existing PDF to clear, and accepted by every institution that accepts PDF/A.

How This Tool Works

Drop your PDF. The tool walks every page’s font table and inspects each FontDescriptor. If every font has an embedded program (FontFile, FontFile2, or FontFile3 stream — covering Type1, TrueType, and CFF respectively), it lets you proceed. If any font is missing its program, it shows you which ones and refuses to write the output. There is no automatic font-embedding fallback — that’s a deliberate choice (see below).

When you click Download as PDF/A, the tool:

1. Loads the original PDF bytes (no re-encoding of pages or images).
2. Embeds the sRGB IEC61966-2.1 ICC profile (3 KB) as a Stream object with /N 3.
3. Adds an OutputIntent dictionary of subtype GTS_PDFA1 referencing that ICC stream.
4. Writes an XMP metadata block declaring pdfaid:part = 1, pdfaid:conformance = B, plus standard Dublin Core fields pulled from the existing document info.
5. Saves with useObjectStreams: false (PDF/A-1 is based on PDF 1.4, which predates object streams).

The result is a file your archive should accept. The output is a candidate PDF/A-1b — for formal compliance certification, run it through veraPDF (the open-source PDF/A validator that the EU’s PREFORMA project sponsors).

Why We Don’t Embed Fonts For You

This is the deliberate limit of any browser-based PDF/A converter. To embed a missing font, you need three things: the original font file in the right format (TTF, OTF, or Type 1), a way to subset it to just the glyphs the document uses, and a way to rewrite every page’s content stream to reference the new embedded program instead of the system font name. The first alone makes browser-side embedding impractical — to “embed Helvetica” you’d need to ship a 200 KB Helvetica file (and you’d be technically illegal, since the Helvetica font is licensed and not freely redistributable). For a document using Cambria, Calibri, or Optima, the font isn’t even installed by default on most operating systems — there’s no source for the bytes.

Acrobat Pro and Ghostscript handle this by either falling back to a metrically-compatible substitute (Helvetica → Liberation Sans) and embedding that, or by rasterising the entire page (text becomes a flat image — searchable text is lost). Both approaches are reasonable for desktop tools that ship their own font library. Neither is reasonable for a browser tool.

So we draw a clear line: we add the metadata required by PDF/A-1b on PDFs whose content already meets PDF/A-1b’s content rules. The tool tells you up front which case you’re in. If you need automatic font embedding, the right tools are:

• Acrobat Pro — File → Save As Other → Archivable PDF (PDF/A). Handles font embedding automatically (substitutes when the original isn’t available).
• Ghostscript (open source) — gs -dPDFA=1 -dPDFACompatibilityPolicy=1 -sDEVICE=pdfwrite -sOutputFile=out.pdf -dPDFSETTINGS=/printer in.pdf. The -dPDFACompatibilityPolicy=1 flag tells Ghostscript to convert non-compliant content to compliant content rather than failing.
• veraPDF — not a converter, but the canonical compliance validator. Run it on the output of any tool to confirm.

What This Tool Will Not Do

• Embed missing fonts. Explained above.
• Strip features that PDF/A forbids. If your source PDF uses transparency, JavaScript, encryption, or external references, the tool doesn’t currently remove them. (You can use Repair PDF for a tolerant re-save first; that strips encryption.)
• Convert PDF/A-2 or PDF/A-3. We target PDF/A-1b only. The other variants need different OutputIntent subtypes and additional XMP metadata.
• Validate the output. The tool produces a PDF/A-1b candidate. For formal compliance, run veraPDF.
• Sign the result. Use Sign PDF online after archiving — chained from the success screen.

Pair With Other Tools

After conversion:

• Sign PDF online — apply a signature to your archival copy.
• Compress PDF — if the source PDF was image-heavy and your archive has size limits.
• Edit PDF metadata — set a meaningful title and author before archiving.

Before conversion:

• Repair PDF — tolerant re-parse (also strips encryption, which PDF/A forbids).
• OCR PDF — if your source is scanned, run OCR first so the archived version has searchable text.

Signegy provides general information, not legal advice. PDF/A compliance requirements vary by jurisdiction and use case — consult your archivist or counsel for specifics.