Pdf Tools

What the pdf-tools MCP server does

The pdf-tools MCP server provides PDF manipulation capabilities for AI agents. The listed operations cover common document changes: combining multiple PDFs, separating pages into smaller documents, changing page orientation, applying watermarks, adding page numbers, updating metadata, and flattening PDFs.

These capabilities fit workflows where an agent needs to prepare or revise an existing PDF. For example, an agent could merge related documents into one file, split a document into page ranges, or apply consistent page annotations. The available description identifies the tool set, but does not specify whether operations run locally or through a hosted API, what file inputs are accepted, or where generated files are returned.

Tools and capabilities

The published capability list includes:

• Merge PDF files.
• Split PDF files.
• Rotate PDF pages or documents.
• Apply watermarks.
• Add page numbers.
• Edit PDF metadata.
• Flatten PDF documents.

The pdf-tools MCP server is therefore oriented toward document transformation rather than PDF content extraction or text analysis. The supplied material does not list OCR, parsing, search, conversion to other formats, form filling, signing, or validation features, so those functions should not be assumed.

How it works

The service is described as providing PDF tools for AI agents through MCP. That means its role is to make the listed PDF operations available to an MCP-compatible agent. Beyond that, the supplied information does not document the tool names, argument schemas, response format, transport, authentication model, or handling of uploaded files.

Because those implementation details are not available here, developers should verify the PDFia developer documentation before building an automated workflow. In particular, confirm how an agent supplies source files, how it selects pages, and how it receives the resulting PDF.

Limitations and notes

The pdf-tools MCP server’s documented scope is limited to the seven PDF operations listed above. No README or configuration reference was provided with this listing. Installation steps, required environment variables, licensing, supported MCP clients, size limits, and file-retention behavior are therefore unspecified.

The description also does not explain whether watermark placement, page-number formatting, rotation angles, metadata fields, or flattening behavior can be customized. Treat those details as implementation questions to resolve from the provider’s documentation rather than as guaranteed features.