PDF accessibility fixer

What it does

It sets the document title and language. The title is the one you type, or one taken from the text if you leave it blank. The language is en-US unless you give another.

If it finds headings, it adds a contents page at the front that lists up to 20 of them. An AI model picks them out of the first 10,000 characters of text, so in a long document it only sees the early ones.

The report says whether the PDF was already tagged, how many images it has, and what else it can see is wrong.

What it does not do

It does not tag the PDF. Tags are the structure that tells a screen reader which text is a heading, what order to read things in and what each image shows, and a WCAG or Section 508 review expects them. So this is not a compliance tool. It sets the title and language and tells you how much is left to do.

It also fails on a document whose headings use characters outside Western European alphabets, such as Cyrillic or Chinese.

How it works

1. Open the tool and upload the PDF, up to 25 MB. Type a title and language if you want to choose them yourself.
2. Pay $1: by card, or with x402 if you’re an AI agent. No signup.
3. Download the new PDF and read the report.

Prefer to do it yourself?

Acrobat Pro sets the title and language in Document Properties, and its accessibility tools add the tags, reading order and alt text this tool leaves out. If you have it, you don’t need this.

Why it’s worth $1

It tells you whether a PDF is tagged and how much a screen reader would miss before you pay for proper remediation, and it sets the title and language while it’s there.

Frequently asked

Will this make my PDF WCAG or Section 508 compliant?

No. Those reviews expect a tagged PDF, and this tool does not add tags. Its report tells you whether the file is tagged and how many images would need alt text once it is.

Will it change how the PDF looks?

Your existing pages stay as they are. If it finds headings, it adds one page at the front that lists them as plain text, with no links or page numbers. It is an ordinary page, not PDF bookmarks, so delete it if you don’t want it.

Does it work on a scanned PDF?

Type the title yourself, since a scan has no text to take one from. It sets that and the language, but it finds no headings, and it can’t make a scan readable: that needs OCR first, which this tool doesn’t do.

Is my PDF kept?

No. The finished PDF comes straight back to you and we don’t keep a copy. To find the title and headings, the first 10,000 characters of its text are sent to OpenRouter, an outside AI service, which runs the model (Llama 3.1 8B) at one of the companies that host it. How long they keep it is up to their policies, not ours. More on the privacy page.