url-to-pdf MCP server

url-to-pdf

Turn any public web page into a PDF, a full-page screenshot, or clean Markdown — with one API key.

No browser farm to run, no headless Chrome to babysit. url-to-pdf renders pages on Cloudflare's infrastructure and hands you a download link. Use it from any MCP-compatible AI client (Claude Desktop, Cursor, and friends) by just talking, or call the HTTP API directly from your scripts.

Why you'll like it

• Four outputs, one key. PDF (A4 / Letter / Legal, portrait or landscape), full-page PNG/JPEG screenshots, clean Markdown (navigation stripped out), and structured JSON (title, description, canonical URL, language, main text, links).
• Your assistant actually sees the result. Small outputs (≤1 MB) are returned inline — screenshots come back as images the model can look at, not just a link it has to read out loud.
• Every render comes with a receipt. Final URL after redirects, HTTP status, capture time, duration, byte size and a SHA-256 of the artifact, so a result can be quoted and verified later.
• Just talk to it. Connect it as an MCP server and say "turn https://example.com into a PDF" — your client does the rest. No curl, no JSON to hand-write.
• Also a plain HTTP API. Every tool is a single POST with your x-api-key. Great for scripts, backends, and low-code platforms.
• Mobile, retina and dark mode. Set any viewport (320–3840 × 240–2160), device pixel ratio up to 3, and render with prefers-color-scheme: dark.
• SSRF-protected. Only public http(s) URLs are accepted; internal and reserved addresses are rejected at the edge.
• Private by design. Outputs are kept for 1 hour then auto-deleted. We never see your card (Paddle handles payments) and your email is used only to recover your key.

Get your API key

1. Go to url-pdf.hammbox.com/buy and complete checkout (Paddle — major credit cards).
2. Your key appears on screen the moment payment completes. Copy it and keep it safe.
3. Lost it? Recover it anytime at /portal/claim with the email you used at checkout. No password, no support ticket.

Use it as an MCP server (recommended for AI clients)

Add url-to-pdf to your client's MCP configuration. With Claude Desktop, Cursor, or any client that supports Streamable HTTP:

{
  "mcpServers": {
    "url-to-pdf": {
      "type": "http",
      "url": "https://url-pdf.hammbox.com/mcp",
      "headers": { "x-api-key": "YOUR_API_KEY" }
    }
  }
}

That's it — no local install. Then just ask your assistant in natural language:

"Convert https://news.ycombinator.com to a PDF and give me the download link."

The client calls the tool, passes your key, and shows you the result.

Use it as an HTTP API (for scripts & backends)

Every render is a POST to a /tools/* endpoint with your key in the header.

# Web page → PDF
curl -X POST https://url-pdf.hammbox.com/tools/url-to-pdf \
  -H "Content-Type: application/json" \
  -H "x-api-key: YOUR_API_KEY" \
  -d '{"url":"https://example.com","format":"A4"}'

# Full-page screenshot
curl -X POST https://url-pdf.hammbox.com/tools/url-to-screenshot \
  -H "Content-Type: application/json" \
  -H "x-api-key: YOUR_API_KEY" \
  -d '{"url":"https://example.com","fullPage":true,"format":"png"}'

# Extract Markdown
curl -X POST https://url-pdf.hammbox.com/tools/url-to-markdown \
  -H "Content-Type: application/json" \
  -H "x-api-key: YOUR_API_KEY" \
  -d '{"url":"https://example.com"}'

# Structured data as JSON
curl -X POST https://url-pdf.hammbox.com/tools/url-to-extract \
  -H "Content-Type: application/json" \
  -H "x-api-key: YOUR_API_KEY" \
  -d '{"url":"https://example.com","maxLinks":50}'

Shortened. The full README is on GitHub.