Check a PDF on a public website for the structure screen readers use.
Maintain this server? Claim it
Call this server's tools from your browser. Nothing to install.
Requests go through mcptry because browsers can't call most MCP servers directly. Header values are sent to the server for that request only and are never stored or logged.
Tool list from our last check. Connect live to refresh it.
Downloads one PDF from a public http or https link and reports the structure that screen readers use: whether the file is tagged, whether it has a document title and a language, how many images have no text description and on which pages, and reading-order signals (pages with columns are flagged for a person to check). Also reports the page count and which pages look scanned or contain tables. Use when the user gives a link to a PDF on a public website and asks about its accessibility or screen-reader structure. Does not open files behind a login, does not change the file, and does not say whether the file meets PDF/UA, WCAG or any law.
Link to one PDF on a public website: http or https, at most 2,048 characters.
When true, adds the page-by-page list.
Pick your client. Values in capitals, like YOUR_TOKEN, are yours to fill in.
claude mcp add --transport http pdf-check https://pdf-check.cybernative.ai/mcp
Downloads one PDF from a public http or https link and reports the structure that screen readers use: whether the file is tagged, whether it has a document title and a language, how many images have no text description and on which pages, and reading-order signals (pages with columns are flagged for a person to check). Also reports the page count and which pages look scanned or contain tables. Use when the user gives a link to a PDF on a public website and asks about its accessibility or screen-reader structure. Does not open files behind a login, does not change the file, and does not say whether the file meets PDF/UA, WCAG or any law.
| Argument | Type | Description |
|---|---|---|
| url* | string | Link to one PDF on a public website: http or https, at most 2,048 characters. |
| detail | boolean | When true, adds the page-by-page list. |
{
"type": "object",
"required": [
"url"
],
"properties": {
"url": {
"type": "string",
"maxLength": 2048,
"description": "Link to one PDF on a public website: http or https, at most 2,048 characters."
},
"detail": {
"type": "boolean",
"default": false,
"description": "When true, adds the page-by-page list."
}
},
"additionalProperties": false
}For one to five links to PDFs on public websites, counts pages by type and returns a price estimate from CyberNative AI LLC's published Terms for two options: an accessible web page made from the PDF (USD 6 per page with tables or scanned pages, USD 4 per other page), and, where the file qualifies, the same PDF with tags added (USD 4 per page, minimum USD 100 per order). Lists files and pages that are not converted, such as fillable forms, maps and drawings, with the reason. The estimate is not binding and nothing is ordered by calling this tool; the price is confirmed by email before any work. Use when the user asks what it would cost to have these PDFs made accessible.
| Argument | Type | Description |
|---|---|---|
| urls* | array | One to five links to PDFs on public websites: http or https, at most 2,048 characters each. |
| detail | boolean | When true, adds the page-by-page count for each file. |
{
"type": "object",
"required": [
"urls"
],
"properties": {
"urls": {
"type": "array",
"items": {
"type": "string",
"maxLength": 2048
},
"maxItems": 5,
"minItems": 1,
"description": "One to five links to PDFs on public websites: http or https, at most 2,048 characters each."
},
"detail": {
"type": "boolean",
"default": false,
"description": "When true, adds the page-by-page count for each file."
}
},
"additionalProperties": false
}We read every tool's name, description and inputs the way the model will, looking for instructions hidden from you. A clean result means no warning signs, not a guarantee. Scanned Nov 1, 2026.