What is PDF/A, and do you need it?
PDF/A is an ISO-standardized form of PDF designed to preserve a document’s static visual appearance over time. You need it when an archive, court, regulator, records policy or submission portal specifically requires PDF/A; for ordinary sharing and short-term use, a standard PDF is usually enough. If you are comparing it with other file types, start with the document formats guide.
How PDF/A makes a document more dependable for long-term preservation
An ordinary PDF can contain everything needed to display a document correctly, but the PDF format does not always require that self-containment. A file may depend on fonts that are unavailable on another computer, use color information that different devices interpret differently, include encrypted content or rely on interactive features that future software may not support.
PDF/A narrows the wider PDF specification by requiring some features and restricting or prohibiting others. Its purpose is to make the document’s static pages reproducible without depending on the software, operating system or computer used to create them.
A conforming PDF/A file normally needs appropriate embedded fonts, defined color information, standardized metadata and an internal declaration identifying its PDF/A part and conformance level. Encryption is not permitted because future readers and preservation systems must be able to access the document without a password or an unavailable security mechanism.
These rules make PDF/A more suitable for finalized records such as contracts, reports, legal filings, invoices, research papers, government records and scanned documents that may need to remain readable for many years. They do not make every PDF/A file a complete preservation package, however. The file still needs suitable storage, backups, integrity checks and a retention policy.
PDF/A is a restricted PDF, not a different-looking file
A PDF/A document normally looks like the ordinary PDF from which it was created. The difference lies mainly in its internal construction and the features it is allowed to use.
| Characteristic | Ordinary PDF | PDF/A |
|---|---|---|
| Primary purpose | General viewing, sharing and document exchange | Long-term preservation of a stable page-based document |
| Fonts | May use embedded fonts or depend on substitution | Fonts and the glyph information needed for rendering must meet the selected profile’s requirements |
| Color | May rely on device-dependent color settings | Color information must be defined so future systems can reproduce it consistently |
| Encryption | Password protection and encryption are allowed | Encryption is prohibited |
| Metadata | Can be limited or inconsistent | Standardized metadata and PDF/A identification are required |
| Interactive content | Can include scripts, multimedia, forms and other dynamic behavior | Interactive features are restricted according to the PDF/A part |
| External dependencies | May refer to content or resources outside the file | Dependencies that threaten reliable future rendering are restricted |
| Validation | Opening successfully may be enough for everyday use | The final file should be validated against a named PDF/A profile |
PDF/A is concerned primarily with preserving the document’s static representation. It does not prove that the information is correct, that every page is searchable or that the file complies with a particular organization’s records policy.
It also does not automatically make a document accessible. Some earlier profiles include requirements for document structure or Unicode text mapping, but modern accessibility normally requires separate attention to tagging, reading order, alternative text and standards such as PDF/UA.
The number identifies the PDF/A part
PDF/A is a family of standards rather than one format with a single set of features. The number after PDF/A- identifies the part of ISO 19005 and the underlying generation of PDF technology.
| PDF/A part | Based on | Practical distinction |
|---|---|---|
| PDF/A-1 | PDF 1.4 | The original 2005 standard. It supports relatively simple archival documents but excludes features such as transparency and file attachments. |
| PDF/A-2 | PDF 1.7 | Adds support for later PDF features, including transparency, layers and JPEG 2000 compression. Attachments are tightly limited. |
| PDF/A-3 | PDF 1.7 | Uses the same general model as PDF/A-2 but permits arbitrary embedded files, such as XML data, spreadsheets or original source documents. |
| PDF/A-4 | PDF 2.0 | The newer generation based on ISO 32000-2. It includes profiles for embedded files and specialized engineering content. |
PDF/A-3 can be useful when a human-readable archival document and its machine-readable source data must travel together. An electronic invoice, for example, might use the visible PDF pages for people and include structured invoice data as an attachment.
PDF/A-4 is the published PDF 2.0-based part of the family. Its base profile is written simply as PDF/A-4, while PDF/A-4f supports arbitrary embedded files and PDF/A-4e covers engineering workflows that can include 3D content. ISO 19005-4:2020 remained the published edition on 2 October 2026, although a revision was under development. A draft revision does not replace the profile named in submission instructions.
The letter identifies the conformance level
PDF/A-1, PDF/A-2 and PDF/A-3 can also carry a lowercase conformance letter:
- Level b, or basic: concentrates on preserving the reliable visual appearance of the pages. PDF/A-1b, PDF/A-2b and PDF/A-3b do not by themselves guarantee searchable text, correct reading order or an accessible document structure.
- Level u, or Unicode: adds requirements for mapping text to Unicode, which supports reliable searching, copying and text extraction. It is available with PDF/A-2 and PDF/A-3.
- Level a: adds structural and semantic requirements to the visual-preservation rules. It is available with PDF/A-1, PDF/A-2 and PDF/A-3, but the letter should not be treated as proof that the document meets every current accessibility requirement.
The profile name matters as a whole. PDF/A-1b, PDF/A-2u and PDF/A-3a are different targets, even though all are members of the PDF/A family. PDF/A-4 does not use the older a, b and u designations.
A normal PDF version number is also not a PDF/A profile. A file described as PDF 1.7 is not automatically PDF/A-2 or PDF/A-3, and a PDF 2.0 file is not automatically PDF/A-4. The file must meet the additional restrictions and identify the relevant PDF/A standard in its metadata.
How to tell which PDF/A requirement applies to your submission
Use PDF/A when the recipient, submission system or retention policy asks for it. Do not convert every PDF to an arbitrary PDF/A profile merely because archival PDF sounds safer.
Instructions may say only “PDF/A,” or they may specify an exact target such as PDF/A-1b, PDF/A-2u or PDF/A-4f. They may add separate conditions covering searchable text, accessibility, signatures, file size, image resolution, attachments or metadata. Those conditions are not interchangeable with PDF/A conformance.
Before preparing the file, check the following:
- Read the complete submission instructions. Look for the exact PDF/A part and conformance level, not just the letters “PDF.”
- Check the portal rather than relying on an old template. An organization may update its accepted profiles while downloaded forms or internal instructions remain unchanged.
- Ask whether attachments are permitted. PDF/A-3 and PDF/A-4f can contain embedded files, but an archive may prohibit them even when it accepts those profiles.
- Confirm whether searchable text is required. A scanned document can satisfy a basic visual profile without useful OCR. A rule requiring a text-searchable PDF needs an additional check.
- Treat accessibility as a separate requirement. A request for an accessible PDF may require PDF/UA, WCAG-related checks or the recipient’s own accessibility rules in addition to PDF/A.
- Check the required signing sequence. Converting or editing a digitally signed PDF can invalidate its signature. Some workflows require conversion and validation before the signature is applied.
- Keep the original source document. PDF/A is usually a final-state representation. It does not necessarily replace the editable word-processing file, spreadsheet, source images or structured data that an archive may also require.
Archival institutions do not all make the same choice. The Library of Congress lists PDF/A among its preferred page-layout formats for several types of textual material. U.S. National Archives guidance permits different PDF/A versions according to project requirements and, in some contexts, specifically limits the use of attachments. A submission can therefore be valid PDF/A yet still fail the recipient’s local rules.
How to create and verify the PDF/A file the recipient will accept
The safest approach is to create PDF/A directly from the original document with software that supports the required profile. Direct export gives the application a better chance of embedding fonts, retaining searchable text, preserving links and transferring document structure correctly.
Printing to a virtual PDF printer may reproduce the visible pages, but it can discard tags, bookmarks, links, metadata, form behavior and other useful information. It should not be the first choice when a proper PDF/A export or conversion function is available.
Convert from the best available source
Use this workflow when you still have the original document:
- Finish the source document first. Resolve comments, tracked changes, hidden sheets, missing images and broken links before conversion.
- Identify the required PDF/A profile. Select the exact part and level named by the recipient. Do not choose PDF/A-1b simply because it appears first in an export menu.
- Check font permissions and availability. A converter may be unable to embed a font if the font is missing, damaged or licensed against embedding. Substitute a suitable font in the source rather than ignoring the warning.
- Export directly to PDF/A. Use an ISO 19005 or PDF/A option in the source application or a dedicated PDF converter.
- Inspect the converted pages. Check equations, special characters, transparency, color, page dimensions, links and images. Conversion can preserve conformance while still changing content in an unacceptable way.
- Run OCR if the source is a scan and searchable text is required. Confirm the recognized text rather than assuming that a successful PDF/A conversion created accurate OCR.
- Validate against the selected profile. Save the validation report when the recipient or records process requires evidence.
- Apply later workflow steps carefully. Merging pages, adding attachments, editing metadata or signing the document can affect conformance, so validate the final submitted file again.
Conversion software may need to modify features that the chosen profile does not permit. A PDF/A-1 conversion might flatten transparency, remove prohibited content or change how layers behave. A converter may also remove encryption because an encrypted file cannot conform to PDF/A.
Reviewing the output is therefore as important as obtaining a technical pass. A valid file with a missing symbol, substituted font or incorrectly flattened diagram is not an acceptable archival copy of the intended document.
Repair and validate an existing PDF
If the editable source is unavailable, use a converter or preflight tool that reports the reasons the PDF fails the selected profile. Common failures include fonts that are not embedded, missing color profiles, prohibited encryption, invalid metadata, unsupported actions and attachments that the target profile does not allow.
Repair the reported problems rather than merely adding PDF/A identification metadata. The internal claim does not make a nonconforming file valid.
A practical repair process is:
- Make a copy of the original PDF and leave the original unchanged.
- Choose the exact validation profile required by the recipient.
- Run the validator and review every failed rule.
- Use a PDF/A conversion or preflight tool to correct the failures.
- Compare the repaired document with the original page by page.
- Check searchable text, reading order and accessibility separately when required.
- Validate the repaired file again.
- Repeat validation after any merge, metadata edit, attachment change or signature operation.
Open-source validators such as veraPDF provide profiles for PDF/A-1, PDF/A-2, PDF/A-3 and PDF/A-4, including their applicable conformance levels. A commercial preflight tool can also be suitable if it supports the exact profile you need.
Different validators can occasionally report edge cases differently because PDF/A depends on several related specifications, including rules for PDF syntax, fonts, color profiles, images and metadata. When a receiving portal rejects a file that passes locally, obtain its error report if possible and test with the profile or validator that the recipient names.
Choose PDF/A because the document has a preservation requirement or because the recipient asks for a particular profile—not because it is automatically better for every PDF. For normal distribution, collaboration or password-protected delivery, an ordinary PDF may preserve useful features that PDF/A must restrict.
Source: PDF/A purpose, parts and conformance requirements from ISO 19005 — PDF/A, PDF Association, and ISO 19005-4:2020 — Document management — Electronic document file format for long-term preservation, International Organization for Standardization; preservation characteristics and institutional use from PDF/A Family, PDF for Long-term Preservation, Library of Congress, and Appendix A: Tables of File Formats, U.S. National Archives and Records Administration; validation profiles and checks from veraPDF Validation, veraPDF. Page last checked 2 October 2026.