PDF Splitter

PDF Splitter is an MCP server that provides random access to PDF contents for selective text extraction, search, and page rendering. It connects AI assistants directly to local PDF files or remote URLs, loading documents into memory so agents can inspect targeted sections rather than parsing entire files at once. Software developers, researchers, and technical analysts use this server to reduce context window load and token consumption when analyzing large manuals, reports, and documentation. By exposing tools to fetch document outlines, search text via regex or plain keywords, extract individual pages, and render high-resolution images or thumbnails, it enables LLMs to process complex documents granularly. Users can also inspect embedded images, extract diagrams directly to disk or base64 format, and batch-render pages at custom DPI settings.

Category: Files, Documents & PDFs

Tags: document processing, extraction, file-manipulation, pdf

Visit PDF Splitter

How to install and configure PDF Splitter

1. Install the Bun runtime if it is not already installed on your system. 2. To install automatically for Claude Code, run: bash bunx espresso3389/pdf-splitter-mcp install claudecode Or for Gemini CLI, run: bash bunx espresso3389/pdf-splitter-mcp install geminicli 3. Alternatively, manually clone the repository and install dependencies: bash cd /path/to/pdf-splitter-mcp bun install 4. Add the server configuration to your MCP client config file: json { "mcpServers": { "pdf-splitter": { "command": "bun", "args": ["run", "/path/to/pdf-splitter-mcp/src/index.ts"] } } } 5. Restart your AI client to load the tools.

What you can do with PDF Splitter

• Extract specific page numbers or page ranges from massive documents to limit LLM context token usage. - Search loaded PDF files using plain text or regular expressions to locate relevant sections and page references. - Retrieve structured document outlines and tables of contents to navigate complex multi-section reports. - Render individual pages or full page batches into PNG or JPEG images at specified DPI levels. - Extract embedded images and figures directly from documents to base64 strings or local file paths.

Key facts

• https://github.com/espresso3389/pdf-splitter-mcp
• Files, Documents & PDFs
• document processing, extraction, file-manipulation, pdf

Part of MCP Servers

How do I install PDF Splitter?

PDF Splitter requires Bun. You can run automated setup commands using bunx espresso3389/pdf-splitter-mcp install claudecode or geminicli, or clone the repository locally, execute bun install, and point your client configuration to the server entrypoint.

What can PDF Splitter do?

The server loads local or remote PDFs into memory and provides tools to extract page text, search text using keywords or regex, extract table-of-contents outlines, retrieve embedded image metadata, extract raw images, and render pages as PNG or JPEG images at chosen DPI resolutions.

Which MCP clients work with PDF Splitter?

PDF Splitter works with any Model Context Protocol client that supports standard command-based servers, including Claude Code, Gemini CLI, and other configurable MCP environments running via Bun.

How does PDF Splitter reduce token usage?

Instead of converting and streaming an entire document into an LLM context window at once, PDF Splitter permits random access so the AI only extracts the specific pages, ranges, or search snippets needed for a task.