Image Tiler is an open-source MCP server that splits large images or screenshots into optimally-sized tiles tailored for major vision LLMs (Claude, ChatGPT, Gemini, etc.). It prevents automatic downscaling and loss of detail by processing images at native resolution, enabling accurate analysis by LLMs. Designed for developers, QA teams, and researchers, it supports local files and URLs, web page capture via headless Chrome, and integrates versioned output for iterative workflows.
Visit Image Tiler's official website for product details and getting started.