Avanquest publishes an open-source MCP (Model Context Protocol) server that gives the Claude Desktop app direct access to PDF API tools — so a user can ask Claude in plain language to convert, merge, split, compress, or watermark a file, and have it happen through the Avanquest PDF API.
Supported operations: converting PDFs to Word/PowerPoint/Excel/plain text/JSON; converting HTML, images, and text files to PDF; merging and splitting PDFs; compressing with configurable quality; deleting, rotating, or resizing pages; adding or removing password protection; watermarking; and analyzing PDFs for tags, forms, signatures, and attachments.
1. Clone the server from GitHub (Avanquest-Software/avanquest-pdfapi-mcp-server), then run npm install and npm run build.
2. Get a free API key from the Developer Portal (see the Getting Started articles earlier in this guide).
3. Add the MCP server's configuration to Claude Desktop's claude_desktop_config.json, including your API key and the path to the built index.js. Claude's UI exposes this file via Settings > Developer > Local MCP servers > Edit config.
4. Install and configure the Filesystem MCP extension as well (Settings > Extensions > Filesystem, or manually alongside the Avanquest server) so Claude can read and write local files during the conversion.
5. Restart Claude Desktop completely so the new configuration takes effect.
Example prompts once configured: "Convert /Users/me/report.pdf to a Word document and save it to /Users/me/report.docx", or "Merge part1.pdf and part2.pdf into combined.pdf."
- Input file size limit for the MCP server: 10 MB per file (lower than the 100 MB limit on the core PDF API).
- Processed results are available to download for 10 minutes, consistent with the rest of the PDF API.
- License: MIT.
Note: On Windows, use the short 8.3 path for the npx command (for example C:\PROGRA~1\nodejs\npx.cmd) to avoid a command-line empty-space parsing error — you can find the correct path by running "where npx" in Command Prompt.