PDF.co — Claude Code MCP server: convert and edit PDFs
The official PDF.co MCP server brings PDF conversion, merging, splitting, OCR and form filling into your agent. Handy when documents are part of the codebase.
claude mcp add pdfco-mcp -e X_API_KEY=<X_API_KEY> -- uvx pdfco-mcpWhat PDF.co does
The server exposes PDF.co tools to an AI assistant. Conversion runs both ways: PDFs become JSON, CSV, text, XLS/XLSX, XML, HTML or images, and documents, spreadsheets, images, web pages, HTML and emails become PDFs. Merging and splitting by page or range sit alongside annotation, form reading and form filling.
Text search supports regular expressions and can return table coordinates. Invoices get parsed by an AI tool, and scanned files go through OCR to become searchable. Passwords can be added or removed, and attachments pulled out of a source file. Local files upload to PDF.co servers, while long jobs report progress by job status.
You need an API key from the PDF.co dashboard and UV installed — the server will not start without it. Configuration takes the command uvx, the argument pdfco-mcp and the X_API_KEY environment variable. Desktop apps sometimes miss your terminal's PATH, so the full path to uvx goes in the command instead. Five Excel converters — excel_to_csv, excel_to_json, excel_to_html, excel_to_txt and excel_to_xml — retire on January 5, 2027; excel_to_pdf stays.
What it helps with
- pull text and tables out of a scanned PDF
- combine several PDFs and images into one file
- split a long PDF by pages or ranges
- fill a PDF form with data from the project
- locate a fragment in a document with a regex
Example prompts for the agent
- Convert
invoice.pdfto CSV and show me the table - Merge
part1.pdfandpart2.pdfinto a single file - Find every mention of “payment” in
contract.pdfand return the coordinates
How to install PDF.co
Platform CLI
claude mcp add pdfco-mcp -e X_API_KEY=<X_API_KEY> -- uvx pdfco-mcpReplace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from pdfco-mcp <command>
Platform CLI
codex mcp add pdfco --env X_API_KEY=<X_API_KEY> -- uvx pdfco-mcpReplace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from pdfco-mcp <command>
Config file
{
"mcpServers": {
"pdfco": {
"command": "uvx",
"args": [
"pdfco-mcp"
],
"env": {
"X_API_KEY": "<X_API_KEY>"
}
}
}
}Replace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from pdfco-mcp <command>
Config file
{
"servers": {
"pdfco": {
"type": "stdio",
"command": "uvx",
"args": [
"pdfco-mcp"
],
"env": {
"X_API_KEY": "<X_API_KEY>"
}
}
}
}Replace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from pdfco-mcp <command>
Platform CLI
gemini mcp add -e X_API_KEY=<X_API_KEY> pdfco uvx pdfco-mcpReplace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from pdfco-mcp <command>
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about PDF.co
How do I install PDF.co in Claude Code?
Run claude mcp add pdfco-mcp -e X_API_KEY=<X_API_KEY> -- uvx pdfco-mcp with your own API key.
What does PDF.co do?
It gives an agent PDF.co tools for converting, merging, splitting, searching, OCR and filling PDFs. Everything runs through the API, with the key passed in X_API_KEY.
What do I need before running the server?
An API key from the PDF.co dashboard and UV installed. The config lists the uvx command, the pdfco-mcp argument and the X_API_KEY variable.
