Convert PDF to PDF/A
PDF to PDF/A rewrites a PDF with Ghostscript in PDF/A-1 or PDF/A-2 mode. It is a conversion, not a certificate: check the result in a PDF/A validator if archival compliance is required.
Drag and drop files here, or click to choose
Accepts PDF · up to 500.0 MB per file · up to 20 files at once
You can also paste a file with Ctrl/Cmd+V, or from the clipboard menu.
PDF to PDF/A runs Ghostscript `pdfwrite` with `-dPDFA=1` or `-dPDFA=2` and `-dPDFACompatibilityPolicy=1`. The file is rewritten as a PDF, not validated against a PDF/A checker and not given an embedded output-intent profile you can rely on in an archive audit.
The `outputFormat` field is reduced to a digit test: if the value contains the character `1`, Ghostscript gets PDFA=1; otherwise it gets PDFA=2. Labels such as `pdfa-3`, `pdfa-3b`, and `pdfx` therefore still produce PDF/A-2, not PDF/A-3 or PDF/X.
This is not OCR and not a color-management suite. Transparency, exotic fonts, and encryption can make Ghostscript fail or drop content. Password-protected files are unlocked for you on this page once you enter the password; API callers run Unlock first.
Open the download in a PDF/A validator you trust if archival compliance is the actual requirement. This page is a Ghostscript conversion, not a certificate.
What a run looked like, on a local server with Ghostscript 10.08.0 and a 4-page text PDF of 51,844 bytes. With `pdfa-1` the result was a PDF 1.4 file of 29,849 bytes whose XMP metadata carried `pdfaid:part='1'` and `pdfaid:conformance='B'`. With `pdfa-3b` and with `pdfx` the result was a PDF 1.7 file of 18,224 bytes whose metadata said `pdfaid:part='2'` and `pdfaid:conformance='B'`. The text was still extractable and the font was embedded as a subset. None of the three results contained an OutputIntent entry. Sizes depend entirely on the input, and the file got smaller here because Ghostscript rewrote it, not because the tool compresses.
That test shows what the tool writes, not what a validator will accept. The identification metadata says the file claims PDF/A-1b or PDF/A-2b; whether it meets the standard is a question for a validator such as veraPDF or the checker your archive or court system names. The site's own quick check, `qpdf --check`, only tests PDF structure and reported the PDF/A-1 test file as valid, which says nothing about PDF/A conformance.
Which option to choose: if the requirement says PDF/A-1, send `pdfa-1`. For anything else, including a requirement that names PDF/A-2, the result is a part 2 file, so you can leave the default. If your requirement names PDF/A-3 or PDF/X, this tool cannot deliver it, even though those labels appear in the list; use a converter that does.
If the file is encrypted, the request fails with HTTP 400 and reason `password_required`: enter the password on the page, or run Unlock first over the API. For a file that Ghostscript cannot process, Repair is the neighboring tool to try on the original before converting again.
Features
- Ghostscript `pdfwrite` with `-dPDFA=1` or `-dPDFA=2`
- Format strings containing `1` select part 1; everything else is part 2
- Not a PDF/A validator and not PDF/X
- Anonymous API at /api/v1/convert/pdf/pdfa
When to use this tool
- Get a Ghostscript PDF/A-1 or PDF/A-2 rewrite as a starting point
- Re-save an office export before a checker you already run
- Avoid claiming PDF/X when you only needed a pdfwrite pass
- Produce a PDF/A-1 candidate for a records system that names part 1, then validate it with the system's own checker
- Re-save a Word export as a part 2 PDF with embedded font subsets before sending it to an archive intake
How do I convert a PDF to PDF/A?
- Upload the PDF; if it is password-protected, enter its password when asked.
- Pick an outputFormat if you care about part 1 vs part 2. Values without a `1` become PDF/A-2.
- Click Process and download the rewritten PDF.
- Run your own PDF/A checker before calling the file archival.
- If the checker rejects it, note the rule it names and fix the source document, then convert the original again instead of converting the output a second time.
- Keep the original PDF. The converted copy was rewritten by Ghostscript, so it is a derivative, not the original.
Limits and edge cases
- No PDF/A-3, no PDF/X, no veraPDF report
- Password-protected files need the password; API callers run Unlock first
- Not OCR; scans remain images
- No OutputIntent entry was present in the test outputs, so do not assume an embedded color profile
- The result is a rewritten copy; output size and appearance can differ from the original
- Upload limit on this website: 500 MB per file, sent in chunks above about 95 MB. A single direct API request body is capped at 100 MB.
Examples
- Choosing pdfa-1 sends -dPDFA=1 to Ghostscript
- Choosing pdfx still sends -dPDFA=2
- A 4-page, 51,844-byte text PDF converted with `pdfa-1` came back as a 29,849-byte PDF 1.4 file with `pdfaid:part='1'` in its metadata
- The same file converted with `pdfa-3b` came back as an 18,224-byte PDF 1.7 file with `pdfaid:part='2'`, identical to the `pdfx` result
- An encrypted copy of that file returned 400 `password_required` and no PDF
Privacy for this tool
Files are uploaded over HTTPS, processed in memory or a short-lived temporary directory on our servers, and deleted when your result is ready. We do not keep copies for later browsing, training, or advertising profiles. See the Privacy Policy for retention details and AdSense cookie disclosures.
Frequently asked questions
- Does pdfa-3 produce PDF/A-3?
- No. Only the presence of the character 1 in outputFormat selects PDFA=1. pdfa-3, pdfa-3b, and pdfx all take the PDFA=2 branch.
- Is the output guaranteed PDF/A?
- No. Ghostscript converts with CompatibilityPolicy=1. Failed subsetting or missing profiles can still yield a file that a veraPDF-class checker rejects.
- Does this embed ICC profiles?
- Not as a documented, guaranteed step. Do not treat the download as a complete output-intent package.
- Will text stay selectable?
- When Ghostscript keeps text as text, yes. Some jobs rasterize or drop content; check the output.
- How do I get PDF/A-1 instead of the default?
- Pick PDF/A-1 (`pdfa-1`). Any outputFormat value that contains the character 1 selects part 1. The default `pdfa` and every value without a 1 select part 2. In a test, `pdfa-1` wrote `pdfaid:part='1'` into the metadata and `pdfa-3b` wrote `pdfaid:part='2'`.
- Why did the PDF/A file come out smaller or larger than my original?
- Ghostscript rewrites the whole file and embeds the fonts it needs as subsets. A 51,844-byte test PDF became 29,849 bytes as PDF/A-1 and 18,224 bytes as part 2. Other inputs can grow, particularly when fonts have to be embedded that were not before.
- Will my archive or court portal accept the result?
- It depends on the portal's checker, which this tool does not run. The output has PDF/A identification metadata for part 1 or 2, level B, but no OutputIntent entry in the test file. Run the portal's validator or your own and treat its report as the answer.
- What happens to a password-protected PDF?
- It is rejected with HTTP 400 and the reason `password_required` until you supply the password on the page, or run Unlock first when using the API.
- Is PDF/A the same as a smaller or safer PDF?
- No. This tool is a Ghostscript rewrite that tags the file as PDF/A part 1 or 2; it makes no promise about size or safety. For a smaller file use Compress.
Last updated:
Call this from code
Every tool on this site is a plain REST endpoint - no account or API key needed for anonymous use. Built for AI agents and developers as much as for browsers.
curl -X POST "https://pdf123.xyz/api/v1/convert/pdf/pdfa" \
-F "[email protected]" \
-F "outputFormat=pdfa" \
-o output.pdfAlso available as an MCP tool for agent clients that speak Model Context Protocol (JSON-RPC 2.0 over POST /mcp). Full API reference
Use it from an AI agent
SkillClaude Code, Codex, Cursor and other AI agents can run PDF To PDF/A for you with this skill: /skills/pdf123-pdf-to-pdfa.md
Files are used only for this processing job and deleted automatically afterward.