MCP Server (AI Integration)

pd comes with a built-in Model Context Protocol (MCP) server, allowing AI agents like Claude Desktop, Cursor, and custom LLM workflows to interact with Pixeldrain.

Starting the MCP Server

Start the stdio-based MCP server with:

pd mcp

Configuration

Claude Desktop

Add the following to your claude_desktop_config.json (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{
  "mcpServers": {
    "pixeldrain": {
      "command": "pd",
      "args": ["mcp"],
      "env": {
        "PIXELDRAIN_API_KEY": "your_api_key_here"
      }
    }
  }
}

Note

PIXELDRAIN_API_KEY is optional. If provided, uploads will be associated with your Pixeldrain account.

Cursor / Other MCP Clients

Configure the command as pd and arguments as ["mcp"].

Available MCP Tools

The pd mcp server exposes the following tools:

Tool	Description
`upload_files`	Uploads one or more base64-encoded files (images, audio, zip archives) and returns a file or album URL.
`upload_text`	Uploads text or markdown content as a file and returns the shareable link.
`download_file`	Downloads a file from Pixeldrain to the specified local path.

With these tools, AI assistants can easily create sharable links for generated artifacts, charts, or images, as well as fetch remote assets during interactive sessions.

Quick Start & CLI End-to-End Encryption