This MCP server provides image processing capabilities via the fetch_images tool. It can fetch images from URLs or load them from local file paths, process them (including handling large images and automatic compression for images over 1MB), and return the results with base64-encoded image data and MIME types. You can run the server using uv and then query the fetch_images tool with a list of image sources. The tool is designed to handle multiple images in parallel, and its outputs include the processed images ready for immediate embedding in clients or further processing in your pipeline.

To use it, start the server in your environment and invoke the fetch_images tool with a list of image sources (URLs or file paths). Examples of usage include fetching a batch of images from the web, loading a local image for transformation, or a mix of both. The tool returns a structured response containing each image as a base64 string along with its corresponding MIME type, enabling straightforward decoding and rendering on the client side.

Prerequisites:

• Python 3.10+
• uv (uv package manager)

Installation steps:

1. Clone the repository: git clone https://github.com/your/repo.git cd repo

2. Create and activate a virtual environment using uv: uv venv

   On Windows:

   .venv\Scripts\activate

   On Unix/MacOS:

   source .venv/bin/activate

3. Install dependencies using uv: uv pip install -r requirements.txt

4. Run the server directly (example): uv run python mcp_image.py

5. Configure for Windsurf/Cursor (optional):

   • Windsurf: add to ~/.codeium/windsurf/mcp_config.json the following entry: { "mcpServers": { "image": { "command": "uv", "args": ["--directory", "/path/to/mcp-image", "run", "mcp_image.py"] } } }

   • Cursor: add a similar configuration under Settings > Features > MCP Servers.

Note: Adjust /path/to/mcp-image to the location of your mcp_image.py and related files.

Tips and common issues:

• Ensure Python 3.10+ is installed and accessible in your environment.
• The server uses UV for execution; if you encounter environment issues, verify that uv is correctly installed and your virtual environment is activated.
• For large images, the server performs automatic compression when the size exceeds 1MB; if you need to customize this threshold, modify the server configuration or code as needed.
• The fetch_images tool accepts a list of image sources (URLs or local paths) and returns an array of results with base64-encoded data and MIME types, making client-side decoding straightforward.
• When using local file paths, ensure the server process has read permissions to those files.
• If you run into connectivity or URL-related errors, verify network access and URL validity. Check server logs for detailed error messages.