PDF Accessibility Inventory

• PDF Accessibility Inventory
• Privacy
• Support

Status: coming soon. The checker is built and tested on example files. It is not open for use, published on Apify or for sale yet, and no price is set.

Technical evidence of observed behavior on the date tested. Not legal advice and not a certification of compliance.

Most sites have more PDFs than anyone remembers. This check finds them and runs named technical checks on each, so a person can start with the ones that matter most.

What it does

You name a site you own or are authorized to test. The checker finds the PDFs on it (the sitemap first, then a small polite crawl of pages on the same host), opens each one and checks:

• whether the file is tagged (has a structure tree),
• the document title and whether it is displayed,
• the language,
• alternative text on figures,
• image-only pages (pages that draw pictures and no text),
• form fields with names, and
• bookmarks for longer files.

Each file gets a priority score whose formula is printed in the report, so the files worth a person's attention first rise to the top. The summary also gives a cost estimate for remediation that multiplies published per-page rates by page count. It is an estimate and a market range, not a quote.

What it cannot tell you

• Automated checks cover only part of PDF/UA and WCAG. Reading order, table structure, color contrast, link text and the quality of alternative text need a person.
• Only PDFs reachable through the sitemap and a capped crawl of the same host are listed. Files behind logins are not.
• The image-only check cannot see text drawn as outlines, and a hidden text layer from earlier text recognition counts as text.
• Its results have not been compared with an independent PDF checker.
• It does not fix files or run text recognition, and it never says a file is accessible.

How you will be able to use it

• An Apify Actor, in development: Apify listing: coming soon One site per run; a maximum number of files per run that you set.
• A printable evidence report is planned and is not available yet.

You confirm that you own the site or are authorized to test it before any fetch; without it nothing is fetched.

How it identifies itself

User agent AgentHarbor-PDFInventory/<version> (+https://agentharbor.agency/pdf-inventory). It reads robots.txt before every URL, waits at least one second between requests to a host (more if the site asks, up to 10 seconds), runs at most two requests at a time, never leaves the host you named, skips files over a size limit, and never logs in, solves CAPTCHAs or works around bot protection. To opt a site out, see our scanner's page.

Waitlist

Paid monitoring and the paid plans for this product do not exist yet, and nothing is for sale. If you want to hear when it opens, e-mail brian@agentharbor.agency with the subject "Waitlist: PDF Accessibility Inventory". We use your address only to tell you about this product and delete it if you ask.

Read the privacy page and the support page.

How this was made: the text and the code behind this product were drafted with an AI assistant (Claude, by Anthropic), then reviewed, tested and run by Agent Harbor LLC.