The merge and page-extract operations are the solid core; the pattern-matching "find related PDFs" is the interesting one, pulling in documents whose text matches a regex from a starting file. Two honesty notes straight from the README: it is marked work-in-progress, and its own author says filesystem or Everything search beats its search — so use this for the manipulation, not as a file finder.
An MCP server that gives a model a handful of PDF manipulation operations — merging files, extracting pages, and locating related PDFs by matching extracted text against a regex. The README's own header marks it work-in-progress and to be used with caution, on Windows.
- Several PDFs merged into one, in default or a specified order
- Specific pages extracted from a PDF
- Related PDFs found — and optionally merged — by matching text against a regex pattern from a target PDF
No account and no key. Install is a clone, then `uv venv`, activate, and `uv pip install -e .`; Claude Desktop launches it with `uv --directory <path> run pdf-tools`. The pyproject names the package `mcp-pdf-tools` at 0.1.0. The README labels the project work-in-progress and says its own search is weaker than a filesystem or Everything search — reach for those first when you just need to locate a file.
Build from source — clone the repository and build it, then point your client at the binary