74 tools · On-device · Local MCP server · Free forever

Everything happens on your device.
Nothing goes outside.

74 PDF tools — merge, convert, compress, OCR, sign, redact — that run in your browser or on your machine, not on our servers. Your files are never uploaded. And for agents and developers, the same tools are a free local MCP server, a local REST API on your own machine, and a Python library.

Private by design — files never leave your device 100% free, forever — no sign-up, no watermark No account, no key, no limits — all 74 tools included No AI training, ever — your documents are never seen by us at all

Browse the tools Connect your AI agent →

Your machine The whole product

Everything pdfmcp does, it does here — on hardware you already own.

  • Browser tools — all 74 tools run in this tab Files open, process and save on this device.
  • Local MCP server — the same 74 tools inside Claude, Cursor and other clients Runs on your machine, talks to your agent locally.
  • Engines, downloaded once — OCR, Office conversion and signing engines Fetched on first use, verified, then cached on your machine for good.
  • Your PDFs stay yours Never uploaded, never logged, never used for AI training.
A cloud server Not used

The usual PDF-site model — the part pdfmcp simply doesn't have.

  • Files uploaded to a server — nothing leaves your machine
  • Accounts, sign-ins and API keys — none exist to create
  • Daily quotas and paid tiers — nothing to run out of, nothing to upgrade
  • Your documents kept or mined — impossible: we never see them

The one exception, stated plainly: an anonymous page-view counter (page path + count only) so we can publish the numbers on /analytics/. It never touches your files.

01Most-loved tools

Local packagePDF to Word (DOCX)Convert a PDF to an editable Word document. Text-based conversion: paragraphs and headings are reconstructed from the text layer; exact layout, tables and images are not preserved.On-deviceMerge PDFsCombine multiple PDF files into one, in upload order.On-deviceCompress PDFReduce file size with Ghostscript. Presets: light (printer quality), medium (ebook), heavy (screen, smallest).On-deviceSplit PDFSplit a PDF by page ranges (ZIP of PDFs) or one PDF per page.On-devicePDF to textExtract plain text, one form-feed separated section per page.On-deviceImages to PDFCombine images (JPG/PNG/WebP/TIFF/GIF) into a single PDF.Local packageOCR — searchable PDFMake a scanned PDF searchable: pages without a text layer are OCR'd with Tesseract and gain an invisible text layer; pages that already have text are passed through and reported as skipped.On-deviceStrip AI metadata & provenanceRemove AI-generation provenance and identifying metadata from a PDF or image (JPEG/PNG/WebP): C2PA Content Credentials, XMP/EXIF/IPTC AI markers (incl. IPTC DigitalSourceType trainedAlgorithmicMedia), generator Software/Creator tags, EXIF GPS and device data. Metadata only — content and pixels are never modified, image pixels are not re-encoded, and pixel-embedded invisible watermarks (e.g. SynthID) are NOT removed: they live in the image data itself, not in metadata. The result reports exactly what was found and removed, by category.

Convert PDF to Word is our most-used tool — followed closely by merging PDFs, compressing PDFs and making scans searchable with OCR.

02Organize PDF

On-deviceMerge PDFsCombine multiple PDF files into one, in upload order.On-deviceSplit PDFSplit a PDF by page ranges (ZIP of PDFs) or one PDF per page.On-deviceRotate pagesRotate pages by 90, 180 or 270 degrees.On-deviceDelete pagesRemove the selected pages from a PDF.On-deviceExtract pagesCreate a new PDF containing only the selected pages.On-deviceReorder pagesRearrange pages into the given 1-based order (a permutation of all pages).

See all organize pdf tools →

03Convert to PDF

On-deviceImages to PDFCombine images (JPG/PNG/WebP/TIFF/GIF) into a single PDF.Local packageHTML to PDFRender an HTML document to PDF. Fully offline: external resources (images, stylesheets, scripts) are never fetched — inline everything you need.Local packageMarkdown to PDFRender a Markdown document (GitHub-flavoured) to a typeset PDF.On-deviceText to PDFRender a plain-text file to a PDF (monospaced, line-wrapped).Local packageOffice document to PDFConvert Word/Excel/PowerPoint/OpenDocument files (docx, xlsx, pptx, odt, ods, odp, doc, xls, ppt, rtf) to PDF via the LibreOffice backend.Local packageEPUB to PDFConvert an EPUB ebook to a typeset PDF (text-focused: chapter text and headings in spine order; embedded images are not rendered).

See all convert to pdf tools →

04Convert from PDF

Local packagePDF to Word (DOCX)Convert a PDF to an editable Word document. Text-based conversion: paragraphs and headings are reconstructed from the text layer; exact layout, tables and images are not preserved.On-devicePDF to textExtract plain text, one form-feed separated section per page.On-devicePDF to imagesRender each page to an image (ZIP).On-devicePDF to MarkdownExtract text as Markdown, with headings inferred from font sizes.On-deviceExtract text as JSONPer-page structured text: blocks with bounding boxes, ready for agents/LLMs.On-devicePDF to HTMLConvert a PDF to a single HTML document, one section per page, preserving positioned text.

See all convert from pdf tools →

05Optimize PDF

On-deviceCompress PDFReduce file size with Ghostscript. Presets: light (printer quality), medium (ebook), heavy (screen, smallest).Local packageRepair PDFRewrite and repair a damaged PDF with qpdf.On-deviceFlatten PDFBake form fields and annotations into the page content.Local packageConvert to PDF/AConvert a PDF to the PDF/A archival standard with Ghostscript.Local packageLinearize PDFRewrite a PDF linearized ('Fast Web View') so first pages display before the whole file downloads.Local packageOptimize images in PDFDownscale and recompress oversized embedded images, and drop duplicate image data.

See all optimize pdf tools →

06Protect & sign

Local packageEncrypt PDFProtect a PDF with a user password (AES-256).Local packageUnlock PDFRemove password protection (the current password is required).On-deviceStrip AI metadata & provenanceRemove AI-generation provenance and identifying metadata from a PDF or image (JPEG/PNG/WebP): C2PA Content Credentials, XMP/EXIF/IPTC AI markers (incl. IPTC DigitalSourceType trainedAlgorithmicMedia), generator Software/Creator tags, EXIF GPS and device data. Metadata only — content and pixels are never modified, image pixels are not re-encoded, and pixel-embedded invisible watermarks (e.g. SynthID) are NOT removed: they live in the image data itself, not in metadata. The result reports exactly what was found and removed, by category.On-deviceSanitize PDFStrip metadata (incl. XMP), JavaScript, embedded files/attachments, document actions and link/file actions — the safe-to-share version of a PDF.On-deviceSign PDF (visual signature)Place a visual signature on a PDF: upload the signature image as the SECOND file, or let the signer name be typeset. A self-signed certificate (SHA-256 fingerprint) is generated and recorded in the document metadata. This is an electronic signature stamp + certificate record — not a qualified/PKI document signature.Local packageSet PDF permissionsEncrypt a PDF with an owner password and fine-grained permissions (print, copy, modify, annotate, ...).

See all protect & sign tools →

07Edit & mark up

On-deviceWatermark PDFStamp diagonal text across every page.On-deviceAdd page numbersStamp page numbers, format 'Page {n} of {total}'.On-deviceAdd header & footerStamp a header and/or footer line on every page. {page} and {total} placeholders are substituted per page.On-deviceBates numberingStamp sequential Bates numbers (prefix + zero-padded counter) on every page.On-deviceFill PDF formFill AcroForm fields by name. Values: strings for text fields, true/false for checkboxes, numbers are stringified.On-deviceRead form fieldsList all AcroForm fields with names, types, current values and locations (JSON).

See all edit & mark up tools →

08Extract & OCR

Local packageOCR — searchable PDFMake a scanned PDF searchable: pages without a text layer are OCR'd with Tesseract and gain an invisible text layer; pages that already have text are passed through and reported as skipped.On-devicePDF info & metadataPage count, page sizes, metadata, encryption status as JSON.Local packageExtract imagesPull all embedded images out of a PDF (ZIP, original formats).Local packageExtract tables (JSON + CSV)Detect tables and return them as tables.json plus one CSV per table (ZIP).On-deviceExtract bookmarks / TOCReturn the document outline as JSON: level, title, target page.On-deviceExtract attachmentsPull all embedded file attachments out of the PDF (ZIP).

See all extract & ocr tools →

09When a job is too heavy for a browser tab, it still stays home

Some tools need desktop-grade machinery — OCR at scale, Office conversion, cryptography. Those run in the local package on your machine, under the same promise: your file never leaves your device. The package starts small: the first time a tool needs an engine with a portable build (qpdf, OCR language data), it downloads that engine once — SHA-256 verified — and caches it locally. Installer-only engines (Ghostscript, Tesseract, LibreOffice) are used automatically when installed, and the tool prints the exact install command when one is missing. No account, no key, no limits — free forever.

Read the free-forever policy · Read the privacy commitments · See exactly what analytics we collect

10Give your AI agent hands: the free local MCP server

Claude, Cursor, VS Code, Windsurf, Codex, Gemini CLI and more can call every pdfmcp operation natively through MCP — running locally, inside the client, on your machine. No key, two minutes of setup:

Same 74 tools, three ways to run them — all on your machine: web tools in your browser, an MCP server inside your AI client, and a local REST API (plus the Python library). No key, no account, no limits — free forever.

{
  "mcpServers": {
    "pdfmcp": {
      "command": "uvx",
      "args": [
        "--from",
        "https://pdfmcp.org/downloads/pdfmcp-4.2.1-py3-none-any.whl",
        "pdfmcp"
      ]
    }
  }
}

Setup guides for 13 MCP clients · Free forever — all 74 tools · Prefer code? The local API quickstart

11Three ways to run pdfmcp — all of them local

There is no hosted version of pdfmcp in any shape. Every way of running the same 74 tools keeps your files on your own machine:

12How pdfmcp compares

Most PDF sites upload your files to their servers and meter you with accounts, daily caps and paid tiers. pdfmcp does the opposite — everything runs on your device, all 74 tools included, free forever, no limits. The honest, feature-by-feature comparisons, including where the other tool is the better pick:

pdfmcp vs iLovePDF · pdfmcp vs Smallpdf · pdfmcp vs IHatePDF · All comparisons

13Questions

Is pdfmcp really free?

Yes — free forever. The web tools are free with no sign-up and no watermark, and the local package includes all 74 tools with no account, no key and no limits. There's no paid tier and nothing to unlock — see the free page for the whole policy.

Where do my files go?

Nowhere. Web tools run in your browser — the file never leaves your device. The local package runs on your machine — the file never leaves your device there either. pdfmcp has no processing server, so there is no copy of your document anywhere else, and nothing is used for AI training.

What is the MCP server?

MCP (Model Context Protocol) lets AI agents — Claude, Cursor, VS Code and others — call tools directly. The pdfmcp package runs as a local MCP server inside your client and exposes all 74 operations, so your agent can merge, OCR, convert and redact PDFs itself — on your machine, with no key.

Is there an API?

Yes — a local one, on your machine. Run pdfmcp serve and the package starts a REST API on 127.0.0.1:8765 (localhost only): submit jobs with a multipart POST to /v1/jobs, poll them, and download the results — from curl, a script, or any language. No API key needed, because it never leaves your computer. The same engine is also a Python library. Full walkthrough: the local API quickstart guide.

Why is it free? What's the catch?

Your device does the work, so the tools cost us almost nothing to give away. No ads, no selling data — we couldn't sell your data anyway, because we never receive it. The only analytics on this site are page-hit counts and Google Analytics, published openly on the analytics page.

Something broke. Who do I tell?

Email [email protected] — a human reads it.

This was free — and your file never left your device

pdfmcp is free because we want it free. If it saved you time, tell a friend: