MCP Florence2, developed by Jkawamoto, is a Python-based server tool that integrates the Florence-2 model for image processing. It connects AI programming assistants and Large Language Models that support the Model Context Protocol. This free license server extends the capabilities of compatible AI clients by providing specialized visual data processing tools. Its distribution primarily occurs through "uvx" from its GitHub repository, with an alternative ".mcpb" bundle available for Claude Desktop.

What does MCP Florence2 connect to?

MCP Florence2 functions as an MCP server, establishing connections with AI programming assistants and Large Language Models that adhere to the Model Context Protocol. It is compatible with several prominent AI clients, including Claude Desktop, Goose CLI, Goose Desktop, LM Studio, and Cursor. This broad compatibility allows various AI environments to access its specialized image processing capabilities, enhancing their ability to interact with visual content.

Core features and tools

The server exposes three primary callable tools for image processing. The ocr tool is designed for extracting text from images or PDF files. The caption tool generates descriptive summaries of image content. Additionally, a process tool is available, which accepts an image source (either a file path or a URL) and a custom prompt to perform specific operations using the Florence-2 model.

Who is it for?

MCP Florence2 is intended for AI programming assistants and developers who are utilizing these assistants. It enables concrete workflows such as extracting text from visual documents, creating concise image descriptions, and executing custom image processing tasks. These capabilities are valuable for automating visual data analysis within AI-driven applications. Skip this if you are not working with AI programming assistants or LLMs that use the Model Context Protocol.

How to get started

No credentials are required. The server communicates via the stdio transport protocol.

Use cases

• Extracting text from images or PDF files using OCR.
• Generating descriptive captions that summarize the content of images.
• Processing an image file with a custom prompt using the Florence-2 model.

MCP Florence2 offers valuable image processing capabilities for AI programming assistants and developers within the Model Context Protocol ecosystem. Its specialized tools for OCR, captioning, and custom image processing extend the functionality of compatible AI clients. This server provides a straightforward way to integrate visual understanding into AI workflows.