Quickstart

Get your AI agent running VaultTools in 3 steps.

1

Install

The interactive setup detects your AI tools and lets you pick a working directory:

npx @vaulttools/mcp init
2

Configure

The init command writes this to your config file automatically. Here's what it looks like: See all config examples →

~/.claude.json

{
  "mcpServers": {
    "vaulttools": {
      "command": "npx",
      "args": [
        "-y",
        "@vaulttools/mcp",
        "--allowed-dir",
        "~/path/to/your/files"
      ]
    }
  }
}
3

First tool call

Ask your AI agent to use a tool.

Your prompt

"Compress ~/path/to/your/files/photo.jpg to reduce its file size."

Tool call (sent by AI)

{
  "name": "vaulttools_image_compress",
  "arguments": {
    "path": "~/path/to/your/files/photo.jpg",
    "quality": 75
  }
}

Tool response

{
  "output_path": "~/path/to/your/files/photo_compressed.jpg",
  "original_size": "2.4 MB",
  "compressed_size": "890 KB",
  "savings_percent": 63.8
}

The file was compressed locally. The AI never saw the image contents, only the file path and the result metadata.