Does pdf_read upload my PDF to a server?

No. The file is read locally on your Mac and the extracted text is processed on-device. Nothing is uploaded to or stored on any cloud server.

Can it read scanned PDFs?

pdf_read extracts the text layer of a PDF. Digitally created PDFs work perfectly. Scanned documents work if they contain an embedded text/OCR layer; image-only scans without text may return little or no content.

Which AI assistants can use pdf_read?

Claude Desktop, Cursor, VS Code (GitHub Copilot), Windsurf and Zed connect locally via stdio. ChatGPT, Claude.ai, Grok and Perplexity connect through the LMCP Cloud Relay connector.

Read & Extract Text From PDFs With AI on Mac

The pdf_read tool lets your AI assistant open a PDF file on your Mac and extract its full text — so the AI can summarize it, answer questions about it, pull out figures, or feed the content into another task. You never upload the file anywhere. Just type something like: "Read ~/Downloads/contract-v3.pdf and tell me the termination clause and notice period." The AI reads the PDF straight off your disk and answers from the real document.

It works on contracts, invoices, research papers, bank statements, scanned reports, ebooks — any PDF you already have on the machine. Because it runs through LMCP, a free native macOS MCP server with 150+ tools, the AI gets actual document text instead of guessing.

Download LMCP to get started.

Which AI agents work?

pdf_read works with every major AI client. Desktop apps auto-configure through a local stdio connection, while web AIs connect securely through the LMCP Cloud Relay connector.

Desktop (local stdio, auto-configured): Claude Desktop, Cursor, VS Code (GitHub Copilot), Windsurf and Zed.
Web (via LMCP Cloud Relay connector): ChatGPT, Claude.ai, Grok and Perplexity.

Whichever you use, the PDF is read locally on your Mac — the AI client only receives the extracted text it asked for.

Automation

pdf_read becomes powerful when chained with other LMCP tools. A single prompt can read a PDF and act on it: extract an invoice total and add it to a Reminders list, summarize a research paper and email the summary to a colleague, pull dates from a contract and create Calendar events, or read a statement and drop figures into an Excel sheet. Because the AI controls the whole pipeline, you skip the copy-paste between apps.

Context

The AI works with your real Mac data, not a generic template. When it reads your PDF, answers reflect the exact wording, numbers and clauses in that document. Combined with Mail, Contacts, Calendar and Finder tools, the assistant understands who sent the file, where it lives, and what it relates to — turning a static document into something it can reason about in your actual context.

Productivity

Reading a 40-page contract or a dense research paper to find one clause can eat 20 minutes. pdf_read lets the AI surface the answer in seconds. Reviewing a folder of invoices, comparing two versions of an agreement, or extracting tables from a report all collapse from manual scrolling into a single natural-language request. For anyone who handles documents daily, this saves hours every week.

Privacy & GDPR

Everything runs locally on your Mac. The PDF is read from your own disk, the text is processed on-device, and no copy is uploaded to or stored on any server. There are no API keys and no cloud middleman holding your documents. This local-first design makes LMCP GDPR-compliant by architecture — your sensitive files never leave the machine.