Reference
PDF statistics you can actually check
The PDF industry repeats a lot of numbers. This page keeps two kinds: the ISO records that define the format, and the limits this site enforces. The standard names link to the catalogue. The product figures come from the same constants the tools use.
ISO catalogue read on 27 October 2026. The product figures are the limits the site enforces.
1. The standards behind every PDF
PDF is not owned by one company any more; it is ISO 32000. Four records matter in daily work, and the page counts below are lifted straight from the ISO catalogue entries, not estimated.
- ISO 32000-1 · PDF 1.7
Published, reviewed and confirmed in 2023
- Published
- July 2008
- Edition
- 1
- Pages
- 747
The first ISO publication of the PDF format, and still the base that PDF/UA-1 references.
- ISO 32000-2 · PDF 2.0
Published, last reviewed and confirmed in 2026
- Published
- December 2020
- Edition
- 2
- Pages
- 986
The current PDF standard. It replaced ISO 32000-2:2017, which was withdrawn.
- ISO 19005-1 · PDF/A-1
Published, reviewed and confirmed in 2026
- Published
- October 2005
- Edition
- 1
- Pages
- 29
Long-term preservation: specifies how to use PDF 1.4 so documents stay readable decades later.
- ISO 14289-1 · PDF/UA-1
Published, reviewed and confirmed in 2025
- Published
- December 2014
- Edition
- 2
- Pages
- 17
Accessibility: specifies how to use ISO 32000-1 to produce documents that assistive technology can read.
One comparison falls out of those records: PDF 2.0 is 239 pages longer than PDF 1.7 — 986 against 747. The format grew. A reader can still open a file written elsewhere.
2. What the profiles are for
PDF/A exists because archives need files that still open in fifty years. ISO 19005-1 restricts PDF 1.4 — no external dependencies, embedded fonts, defined colour — so nothing outside the file can disappear. Later parts extend the profile to newer PDF versions.
PDF/UA exists because a scanned page is a picture, and a picture of text is unusable with a screen reader. ISO 14289-1 specifies the structure a PDF needs — tagged content, reading order, language, alternative text — before assistive technology can make sense of it.
Both are conformance profiles, not features you can bolt on at the end. A document either validates against the profile or it does not — there is no partial credit, and a page that only looks right is not a conforming file.
PDF to PDF/A makes an archive file. Accessible PDF Check reports tags, language, title, and image text. It does not rebuild a fully tagged PDF.
3. Our own numbers
These are the limits the site enforces. They are not a market estimate, and they are not frozen on the ISO date above.
PDF tools live on the site
53
Counted from the tool list this page is built with.
Languages the interface ships in
18
Each language has its own message pack. English is the source.
Largest everyday file
100 MB
The cap for tools that run on your device. Free conversion and OCR are smaller, and that cap is written on the tool. Pro uses the same 200 MB file ceiling.
Free everyday jobs per day
10
A guest is counted in that browser. A signed-in account is counted once. People who are not on Pro also share a daily ceiling on the same network. The count resets daily.
Free pages per day
500
Shared across everyday tools, then reset daily.
Free advanced uses per day
3
OCR, Office conversion, and AI draw from this.
Pro everyday jobs per day
500
The Pro ceiling for everyday tools.
Pro advanced uses per day
50
The Pro ceiling for OCR, Office conversion, and AI.
Pro pages per cloud job
200
The ceiling for one OCR, Office conversion, or AI job.
4. Numbers we refuse to publish
A statistics page is only worth reading if it leaves things out. These three claims are everywhere, and none of them survives a source check.
“Trillions of PDFs are created every year”
No primary dataset counts annual PDF creation. The figure circulates in marketing pages without an origin, so it does not appear here.
“X% of business documents are PDF”
Vendor surveys measure their own customers, not the world. Without the sampling method published, the percentage is not a statistic.
“PDF is the most-used document format on the web”
Plausible, but proving it needs a crawl dataset with a date and a method. We would rather link the dataset than paraphrase a claim about it.
5. FAQ
Why does this page exist rather than a page of market-size numbers?
Because most PDF “statistics” pages repeat figures with no traceable source. It is more useful to publish what can be verified — the ISO records behind the format, and the limits our own code enforces — than a table of numbers nobody can check.
What is the difference between PDF, PDF/A and PDF/UA?
PDF is the format itself, standardised as ISO 32000. PDF/A (ISO 19005) is a preservation profile: it restricts the format so a file stays readable long-term. PDF/UA (ISO 14289) is an accessibility profile: it specifies the structure assistive technology needs. A file can be one, both, or neither.
How big is the PDF specification?
The ISO listing for PDF 1.7 runs to 747 pages and PDF 2.0 to 986 pages. That is the reason a PDF library is never a weekend project and why conformance is measured, not assumed.
Are the platform numbers on this page marketing figures?
No. The product figures are the same limits the site enforces: daily jobs, the everyday file ceiling, and the Pro ceilings. The ISO rows are separate. Their catalogue date is printed under the title.
Sources
- ISO catalogue · ISO 32000-1:July 2008
- ISO catalogue · ISO 32000-2:December 2020
- ISO catalogue · ISO 19005-1:October 2005
- ISO catalogue · ISO 14289-1:December 2014
- PDFTrusted plans and enforced limits
The ISO records were read on 27 October 2026. Edition, publication month, and page counts are quoted from those catalogue entries. Nothing on this page estimates a market size.