Describe a PDF’s images
Write alt text for a document’s pictures, into the document, where a screen reader reads it.
What this does
A screen reader reading a PDF reads its structure tree, not its page, so a photograph with nothing said about it is silence — and a chart carrying the whole point of a report is silence with a caption underneath. This writes the descriptions into the file itself, on each picture’s own element, so the document gets better rather than you getting some text. The person it helps is not you: it is whoever you send the file to, on a machine this product will never see.
What it will not do
Every tool here says what it cannot do, in its own words, before you rely on it.
- This is the one AI tool that sends pictures rather than words. The pictures going are shown to you before they go, like everything else.
- The descriptions are written into the file, on the picture’s own element. They are not drawn on the page.
- A scan stored as fax or JBIG2, and a JPEG 2000, cannot be handed to a provider and are reported rather than sent.
- A picture drawn from inside a form — a stamp, a repeated header — is described only where the document is already tagged for it.
- Alt text is one part of an accessible document. Reading order, headings, table headers and the document’s language are untouched.
- What the description says is what a model saw. Read it before you rely on it, particularly for a chart whose numbers matter.
Questions people ask
Where does the description end up?
Inside the file, on the structure element that owns the picture — the place every screen reader made in the last twenty years looks. It travels with the document to everybody who opens it.
Does this make my document accessible?
It fixes one part of it, and an important one. Reading order, headings, table headers and the document’s language are separate questions this does not touch.
What exactly is sent to the provider?
The text on the screen under “what will be sent”, and nothing else of your document. It is not a summary of the request — it is the request. Nothing is sent until you press send, and the page cannot send it before showing it to you, because the code refuses.
What does this cost?
Pagecraft charges nothing and never will. Your provider charges you for what you send it, at their rates, on your own account — and you are shown the token count and an estimate in dollars before you send anything. A model on your own machine costs nothing at all.
Do I have to pay for an API key?
No. Install Ollama or LM Studio, run a model on your own computer, and point Pagecraft at it: no key, no bill, and your document never leaves the device. That path is first-class here rather than a degraded fallback.