How do I check every PDF on my website for accessibility?
Find the PDF URLs, send each one to the free PDF Checkup check API at its rate limit, and read each report. Limits and field names on this page were read from the live API on 2026-11-20.
or drop it here. Free: up to 30 pages and 20 MB, no sign-up. The file is deleted as soon as the check finishes.
Accessibility report
Why did my PDF fail? The common failures in plain English
Technical PDF/UA-1 results (veraPDF)
Fix this PDF
After the fix
What are the steps to check all the PDFs on a website?
Make a list of the PDF URLs on your site, send each URL to the free check API at POST /api/v1/check, and read the report for each one. The API needs no account and no key, and it checks one PDF per request.
- Find the PDFs: from your sitemap, a crawl or your content system.
- Send each public PDF URL to the free check API, no faster than the anonymous rate limit of 3 checks per minute per IP.
- Read each report: the summary score, the failed checks and the pages they affect, then fix the worst files first.
How do I find all the PDFs on my website?
Start from the places that already list your files, and merge the results into one file with one public PDF URL per line. No single source catches every PDF, so combine two or more.
- Sitemap: if your XML sitemap lists PDF files, the command below pulls them out. Many sitemaps list only HTML pages, so an empty result means “not listed”, not “no PDFs”.
- Search: the Google query site:example.com filetype:pdf lists the PDFs Google has indexed on a domain you own.
- Content system or crawler: export the media library of your CMS, or crawl your own site with a crawler and keep the URLs that end in .pdf.
curl -s https://example.com/sitemap.xml | grep -oE 'https?://[^<]+\.pdf' | sort -u > pdf-urls.txtHow do I send each PDF URL to the free check API?
POST a JSON body {"url": "<public PDF URL>"} with Content-Type application/json to https://pdfcheckup.com/api/v1/check; the response is a JSON report. The script below reads pdf-urls.txt, checks one URL every 21 seconds and writes one summary line per PDF to results.jsonl (it needs curl and jq).
An example run on 2026-11-20 against two public IRS forms and one broken link returned pages 6, score 67, failed 3 and pdfua "fail" for IRS Form W-9, and the error fetch_failed for the broken link, so failed downloads show up in the results instead of stopping the run.
To send a file from disk instead, POST the PDF itself with Content-Type application/pdf and --data-binary @file.pdf.
while read -r url; do
jq -n --arg u "$url" '{url: $u}' \
| curl -s -X POST -H "Content-Type: application/json" -d @- https://pdfcheckup.com/api/v1/check \
| jq -c --arg u "$url" '{url: $u, pages: .doc.pages, score: .summary.score, failed: .summary.failed, pdfua: .summary.pdfua, error: .error}'
sleep 21
done < pdf-urls.txt > results.jsonlWhat are the limits when I check many PDFs?
Anonymous use is limited to 3 checks per minute per IP address, and each PDF can be up to 30 pages and 20 MB. At 3 per minute, 100 PDFs take about 35 minutes (calculated, not measured).
| Limit | Value | What happens beyond it |
|---|---|---|
| Rate, anonymous | 3 checks per minute per IP | HTTP 429 with a Retry-After header of 60 seconds |
| File size | 30 pages and 20 MB per PDF | HTTP 413; the file is not checked |
| Time per check | 5 to 30 seconds typical | HTTP 503 with a Retry-After header while the checker is busy or waking up; retry |
| API keys | A key raises the rate to 30 per minute, but keys are not offered publicly yet | An invalid key returns HTTP 401; send no key to use the anonymous limit |
| Files | Deleted right after the check; never used to train AI | Not applicable |
Which websites can I check this way?
Use the bulk method on websites you own or are allowed to audit. The PDF Checkup agent terms say not to use the API to crawl third-party sites in bulk.
The PDFs must be public: the checker downloads each URL itself, so a link behind a login or a firewall returns an error such as fetch_failed.
How do I read the report for each PDF?
Read summary.score, summary.failed and summary.pdfua first, then the checks array for the details. summary.score is the share of applicable checks passed, from 0 to 100, and summary.pdfua is pass, fail or unknown for the PDF/UA-1 rules run with veraPDF.
- doc: pages, bytes, title, lang and whether the file is tagged.
- summary: score, passed, failed, warnings, and counts by fix type (autoFixable, aiFixable, manual).
- checks: one entry per check with id, status (fail, warn, pass or na), title, detail, the WCAG success criteria in wcag, the pages affected and fix (none, auto, ai, manual or source).
- Sort results.jsonl by score to find the worst PDFs, and group the failed checks by id to see which problem is most common across the site.
Can I check a few PDFs by hand instead of with the API?
Yes: for a handful of files, upload each PDF to the checker on this page and read the report in the browser. It is free, needs no account and has the same limits of 30 pages and 20 MB per file.
The browser checker takes one file at a time, so it stops being practical once you have dozens of PDFs. Use the API script above for those.
What does a bulk check not tell me?
A bulk check lists the machine-checkable problems in each PDF; it does not show that any PDF is accessible or compliant. Reading order, alt text quality, contrast and link wording need a person, and this is not legal advice.
Fixes are not part of the API. After you find the failing PDFs, open each in the web checker, where the paid page packs fix titles, language, link descriptions, tab order and AI-suggested alt text.
How much does PDF Checkup cost?
PDF Checkup is free to check a PDF of up to 30 pages and 20 MB, and fixing a PDF costs $29 for a 50-page pack or $99 for a 300-page pack, paid once and valid 12 months.
Free checker
$0
- WCAG 2.1 AA and PDF/UA-1 report on screen
- Issues by page, in plain English
- Up to 30 pages and 20 MB per file
- No account, files deleted after the check
Remediation packs
- Automatic fixes: title, language, link descriptions, tab order, tagged flag
- AI-suggested alt text for images, reviewed by you before it is written into the file
- Downloadable compliance report PDF (WCAG 2.1 AA and PDF/UA-1 results; the report itself passes PDF/UA-1)
- Files up to 500 pages and 60 MB. One credit per page, charged once per document
- No subscription. 14-day refund if you change your mind
Frequently asked questions
Is there a limit to how many PDFs I can check?
There is no cap on the number of PDFs, but the free API allows 3 checks per minute per IP address, so 100 PDFs take about 35 minutes at that rate. Space the requests 21 seconds apart, and if you get HTTP 429, wait the 60 seconds in the Retry-After header.
Do I need an API key to check PDFs in bulk?
No. The check API works without an account or key. API keys exist, but PDF Checkup says on its agents page that they are not offered publicly yet.
What happens to a PDF longer than 30 pages?
The free API refuses it with HTTP 413 and does not check it. The web checker accepts files up to 500 pages and 60 MB with a paid page pack.
Can I check PDFs on a website I do not own?
Not in bulk. The agent terms say not to use the API to crawl third-party sites in bulk. Checking a single public PDF by its URL is what the API is for.
Are the PDFs I send by URL stored?
No. PDFs fetched from a URL are deleted right after the check and never used to train AI. API calls are counted by page count only; file names, URLs and contents are not stored.
Does passing this checker make my PDF ADA or WCAG compliant?
No automated tool can prove that. Machine checks catch missing tags, titles, language, alt text and many PDF/UA errors, but a person still has to confirm reading order, alt text quality, colour contrast and meaningful link text. Use the report as your to-do list and keep a human review step. This site does not give legal advice.