MCP Image Extractor

MCP Image Extractor provides tools to extract and convert images to base64 for LLM analysis. You can pull images from local files, fetch them from URLs, or process existing base64 data, all to prepare compact, context-friendly image inputs for AI systems.

How to use

You interact with the MCP Image Extractor through an MCP client by selecting one of the available tools and providing the required input. The tools convert images into base64 data, automatically resizing images to a maximum of 512x512 pixels to keep the base64 output compact and efficient for analysis. Use this to analyze screenshots or images embedded in local projects or web content.

How to install

Prerequisites you need on your system are Node.js and npm (Node package manager). Ensure you have a working Node environment before proceeding.

{
  "mcpServers": {
    "image-extractor": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-image-extractor"
      ]
    }
  }
}

If you prefer a local installation, you can point to the built files after cloning the project. Add this to your MCP client configuration as a stdio server pointing to the built index file.

{
  "mcpServers": {
    "image-extractor": {
      "command": "node",
      "args": ["/full/path/to/mcp-image-extractor/dist/index.js"],
      "disabled": false
    }
  }
}

If you choose to install the package globally and run it as a CLI, perform the manual steps to clone, install dependencies, build, and link, then configure the client to use the CLI command.

# Clone and install
git clone https://github.com/ifmelate/mcp-image-extractor.git
cd mcp-image-extractor
npm install
npm run build
npm link

Then configure in your MCP client:

{
  "mcpServers": {
    "image-extractor": {
      "command": "mcp-image-extractor",
      "disabled": false
    }
  }
}
`

If you run into Cursor-specific issues like a “Failed to create client” error, try the local path installation method or verify you’re using the correct path to the executable.

Additional configuration and notes

Available tools are designed to handle three input scenarios and convert images to base64 with a 512x512 resize to optimize context window usage for large language models.

• extract_image_from_file: Takes a local file path and returns a base64 image string. The image is resized to 512x512 as needed.

• extract_image_from_url: Takes an image URL and returns a base64 image string. The image is resized to 512x512 as needed.

• extract_image_from_base64: Processes a base64-encoded image with an optional MIME type (default: image/png). The image is resized to 512x512 as needed.

Docker

If you prefer containerized usage, you can build and run a Docker image for the MCP Image Extractor.

docker build -t mcp-image-extractor .
docker run -p 8000:8000 mcp-image-extractor

Available tools

extract_image_from_file

Extracts an image from a local file and converts it to base64. Images are resized to a maximum of 512x512 for efficient base64 output.

extract_image_from_url

Fetches an image from a URL and converts it to base64. Images are resized to a maximum of 512x512 for efficient base64 output.

extract_image_from_base64

Processes a base64-encoded image for analysis, with an optional MIME type (default image/png). Images are resized to a maximum of 512x512 for efficient base64 output.