PDF accessibility quick check
Pick a PDF and see what the file says about itself: whether it is tagged, titled and has a language, whether its pages carry text or only a scanned picture, and whether its form fields have names. It reads the file in your browser; nothing is uploaded. It is a first look, not a PDF/UA validator.
Check one PDF
PDF accessibility quick check — altstamp.com
Your PDF is read by this browser, on this device. It is not uploaded or sent anywhere. Pre-filled with a tagged sample menu built in your browser.
Or drop a PDF on this box. Up to 60 MB.
Or try a sample
What the file shows
Passing these checks does not make a PDF accessible
They read what the file says about itself. Whether the tags, reading order and alt text are right takes a full check in PAC and a person with a screen reader. The steps are in how to make a PDF accessible.
What WCAG asks
- 1.3.1Info and RelationshipsA
- 2.4.2Page TitledA
- 3.1.1Language of PageA
- 4.1.2Name, Role, ValueA
- 1.4.5Images of TextAA
What it checks
The check opens the PDF the way a PDF reader does, follows the file's own cross-references to the document catalog, and reports eleven things a file says about itself. It unpacks compressed parts with the decompression built into your browser, so nothing is installed and nothing is sent. Each row ends in a word: Found, Not found or Could not tell.
Scroll sideways to see all columns.
| Row | What it reads | Criterion |
|---|---|---|
| Tags | MarkInfo Marked is true and a StructTreeRoot exists in the catalog | 1.3.1 |
| Document title | Title in the document information, or dc:title in the XMP metadata | 2.4.2 |
| Title in the title bar | DisplayDocTitle true in the viewer preferences (W3C PDF18) | 2.4.2 |
| Document language | Lang in the catalog (W3C PDF16) | 3.1.1 |
| Text on the pages | Text-showing operators (Tj, TJ) in each page's drawing instructions, or only images | 1.4.5 |
| Heading tags | H and H1 to H6 elements in the structure tree, after the role map | 1.3.1 |
| Figure tags without alt text | Figure elements with no Alt or ActualText | 1.1.1 |
| Form fields | AcroForm fields, and which have no tooltip (TU) name | 4.1.2 |
| Bookmarks | An outline with at least one entry | 2.4.5 |
| PDF/UA identifier | pdfuaid:part in the XMP metadata | ISO 14289-1 |
| Blocking permissions | Encryption with permission bit 10 (extraction for accessibility) off | — |
"Could not tell" is an answer, not a failure of the file. It appears when the part that holds the answer is encrypted, or packed in a way this check does not unpack. The check never fills that gap with a guess; open the file in a full checker instead.
How to use it
- Choose a PDF, or drop one on the box. The result appears in a second or two; a long file takes longer.
- Read the rows marked needs work first, starting with tags, title, language and pages with no text. Those fixes matter most.
- For each, the guide shows where to fix it, in the source file or in the PDF.
- Fix, export again, and check again. Then run a full PDF/UA check and a screen reader pass.
The two samples show both ends: a tagged one-page menu that passes these rows, and a scanned sign-up sheet that fails nearly all of them. Use them to see what each word looks like before you try your own file. The fixes, in order, are in how to make a PDF accessible.
What WCAG asks
The criterion behind the first row
Information, structure, and relationships conveyed through presentation can be programmatically determined or are available in text.
1.3.1 Info and Relationships, Level A. In a PDF the structure is carried by tags, which is why the first row of the check is whether the file has any.
W3C's PDF techniques (PDF16 for the language, PDF18 for the title, PDF7 for scans) are examples of ways to meet WCAG, not requirements: what counts is the success criterion. PDF/UA-1 is ISO 14289-1, the standard for the file itself, and its requirements are consistent with WCAG 2.0. Which WCAG version you answer to depends on who you are; 508 compliant PDF covers the federal case.
What a fail usually looks like
- Tags Not found on a file made in Word. It was printed to PDF, or exported with the accessibility tags option off. Export it again; see the export settings in accessible Word documents.
- Text on 0 of 2 pages. A scan. It needs OCR and then tags, or a digital original exported again.
- Title Found, title bar Not found. The commonest near miss: the title is set, but the viewer still shows the file name.
- Fields without a tooltip name. Typical of a form drawn in a design program. A screen reader may announce only "edit text", with no clue what to type.
- A PDF/UA identifier on a file with no tags. The claim is metadata anyone can write. The check reports both, and the tags row is the one that matters.
What it cannot see
It reads what the file declares, not whether the declaration is true. A file can be tagged with every paragraph marked as a heading, carry alt text that says "image", or read its columns in the wrong order, and still pass every row here. Reading order, tag accuracy, table headers, contrast and the quality of alt text need a full checker and a person. PAC (PDF Accessibility Checker) is a free checker published by axes4. It checks many PDF/UA and WCAG requirements automatically; a visual check by a person is still needed. It runs on Windows 10 and 11 only; other checkers are compared in PDF remediation software compared.
Encrypted files
An encrypted PDF keeps its text, title and compressed parts scrambled, so most rows read Could not tell. The permissions row still works: it reads the file's permission flags, which are not encrypted.
If the file needs more work than you have time for, services that hand back tagged, checked files are compared in PDF remediation services compared.