pdfspark-mcp

PDFSpark MCP Server

Convert HTML and webpages to PDF directly from AI assistants like Claude, Cursor, and VS Code.

PDFSpark MCP Server connects AI tools to the PDFSpark API using the Model Context Protocol. Generate, merge, split, and inspect PDFs without leaving your AI workflow.

Quick Start

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "pdfspark": {
      "command": "npx",
      "args": ["-y", "pdfspark-mcp"]
    }
  }
}

Then ask Claude things like:

• "Convert this HTML report to a PDF with landscape orientation"
• "Save https://example.com as a PDF with A3 paper size"
• "Merge these two PDFs into one document"
• "Extract pages 1-3 and 7 from this PDF"
• "How many pages does this PDF have?"

Available Tools

Tool	Description	Key Parameters
html_to_pdf	Convert HTML content to PDF	html, format, landscape, scale, margin, watermark, password
url_to_pdf	Convert a webpage URL to PDF	url, format, landscape, scale, margin, watermark, password
merge_pdfs	Merge 2-100 PDFs into one	pdfs (array of base64 strings)
split_pdf	Extract specific pages from a PDF	pdf (base64), pages (e.g. "1-3,5,8-10")
pdf_info	Get metadata, page count, and dimensions	pdf (base64)

PDF Options (html_to_pdf & url_to_pdf)

Both conversion tools support these optional parameters:

Option	Type	Description
format	A4 | A3 | A5 | Letter | Legal | Tabloid	Paper format (default: A4)
landscape	boolean	Landscape orientation
printBackground	boolean	Include background graphics
scale	number (0.1-2)	Scale factor
margin	object	Page margins in CSS units (top, right, bottom, left)
emulateMediaType	screen | print	CSS media type emulation
metadata	object	PDF document metadata (title, author, subject, keywords)
watermark	object	Diagonal text watermark (text, fontSize, opacity, rotation, color)
outline	boolean	Auto-generate bookmarks from HTML headings
password	object	AES-256/RC4 encryption (userPassword, ownerPassword, permissions)

Architecture

┌─────────────────┐     stdio      ┌──────────────────┐     HTTPS     ┌─────────────────┐
│  Claude/Cursor   │◄──────────────►│  pdfspark-mcp    │──────────────►│  pdfspark.dev   │
│  (AI Assistant)  │                │  (MCP Server)    │               │  (PDF API)      │
└─────────────────┘                └──────────────────┘               └─────────────────┘

Configuration

Environment Variable	Default	Description
PDFSPARK_BASE_URL	https://pdfspark.dev	API base URL

To use a custom API endpoint:

{
  "mcpServers": {
    "pdfspark": {
      "command": "npx",
      "args": ["-y", "pdfspark-mcp"],
      "env": {
        "PDFSPARK_BASE_URL": "https://your-instance.example.com"
      }
    }
  }
}

SoftVoyagers Ecosystem

PDFSpark is part of the SoftVoyagers free API portfolio:

Product	Domain	Description
LinkMeta	linkmeta.dev	URL metadata extraction API
PageShot	pageshot.site	Screenshot & webpage capture API
PDFSpark	pdfspark.dev	HTML/URL to PDF conversion API
OGForge	ogforge.dev	Open Graph image generator API
LinkShrink	linkshrink.dev	Privacy-first URL shortener API
Faktuj	faktuj.pl	Polish invoice generator
QRMint	qrmint.dev	Styled QR code generator API
PageDrop	pagedrop.dev	Instant HTML hosting API

License

MIT