Back to MCP Directory
PDF Tools
Model Context Protocol moderate risk

PDF Tools

A versatile MCP server for PDF manipulation, offering local and remote operations like rendering, merging, metadata extraction, and content extraction.

Connections & Capabilities

Connects To

GitHub

Capabilities

readwrite

Quickstart

Install

npx -y @smithery/cli install @danielkennedy1/pdf-tools-mcp --client claude

Exposed MCP Tools (7)

safe
Display

Renders PDF pages as images.

Read-only operation; no modification of data.

moderate
Merge

Combines multiple pages into a single long page.

Creates a new PDF, but does not modify existing files.

safe
Metadata

Extracts document metadata from a PDF.

Read-only operation; no modification of data.

safe
Text

Extracts text blocks and detailed text information from a PDF.

Read-only operation; no modification of data.

moderate
Snippets

Creates freeform or full-width snippets from PDF pages.

Creates a new PDF snippet, but does not modify existing files.

moderate
Fuse

Combines pages from multiple documents into a single document.

Creates a new PDF, but does not modify existing files.

moderate
Download

Fetches PDFs from URLs to local storage.

Downloads files from the internet; potential for malicious content.

Safety Assessment

This server provides useful PDF manipulation tools, but the write operations and remote fetching capabilities introduce moderate risks. It's reasonably safe for local, trusted PDFs, but caution is advised when processing PDFs from untrusted remote sources. Input validation and resource limits are important considerations.

  • PDF documents are stored with UUID4 filenames for security.
  • Includes both read and write operations, allowing for document manipulation.
  • Clear separation of local and remote operations.
  • Uses well-established libraries like PyMuPDF.
  • Write operations on PDFs could potentially introduce vulnerabilities if not handled carefully.
  • Remote PDF operations involve fetching data from URLs, posing a risk of malicious content.
  • Lack of explicit input validation details in the provided documentation.
  • No mention of resource limits or rate limiting for remote operations.