Convert PDF to PDF/A

PDF/A is the ISO-standardised archival version of PDF (ISO 19005). Its goal is that a document looks exactly the same decades from now, in any viewer. So the file must contain everything needed to display it, such as fonts and a colour definition, and must not contain anything unpredictable, like JavaScript, external content or encryption.

Document management systems, court and tender portals, public archives and accounting software often require PDF/A. The receiving side decides which level you need.

Level	Choose it when
PDF/A-2b	The most common requirement. Allows transparency and modern compression.
PDF/A-1b	Older systems ask for it. Stricter: no transparency, layers or JPEG 2000.
PDF/A-3b	Attached files must be kept, such as the XML of an e-invoice.

The “b” stands for visual preservation: the appearance is guaranteed. Levels “a” and “u” also require verified text structure and Unicode mapping, which this tool doesn't produce.

The converter walks through the PDF's structure and applies the required fixes:

• writes XMP metadata with the PDF/A identification, kept in sync with title, author and dates;
• embeds an sRGB colour profile (OutputIntent);
• removes JavaScript, launch and automatic actions, multimedia, XFA forms and encryption;
• removes attachments for 1b and 2b, or keeps them with the required descriptions for 3b;
• makes annotations printable and strips device-specific settings and image interpolation;
• converts CMYK colours to RGB where it can be done without flattening the page;
• checks that every font is embedded, and replaces missing Arial or Helvetica with a metric-compatible embedded font so text stays put.

If a page still has something that can't be fixed, such as an unusual non-embedded font or transparency in PDF/A-1b, the page is turned into an image with an invisible, searchable text layer. The checklist names those pages, and you can switch this off.

The tool applies the structural rules of the PDF/A standard, and in our tests the output met every rule we checked: embedded fonts, matching metadata, colour profile and no forbidden elements. It is not a certified validation, though. Some fine-grained rules, such as the internals of font programs, are only examined by a full validator.

If a portal rejects the file, its message usually names the rule. A common cause is a digital signature: converting changes the file, so the old signature becomes invalid. Sign after converting. You can inspect signatures with the signature checker.

If a PDF only has permission restrictions (an owner password), encryption is removed automatically, since PDF/A can't be encrypted. If it needs a password to open, remove it first with Remove password. Minor structural damage and forbidden LZW compression are repaired.

Everything runs in your browser. The PDF isn't uploaded or stored, and no watermark is added.

Need to change the title or author? Enter the title before converting rather than editing the metadata afterwards, because changing it later can break the match between the XMP metadata and the document information that PDF/A requires. If the source is a Word or other office document, saving it as PDF/A in that program is often the cleanest route; use this tool for PDFs you already have.