Image Recognition MCP Server

A Model Context Protocol (MCP) server that provides AI-powered image recognition and description capabilities using OpenAI-compatible vision models.

Overview

This MCP server enables AI assistants to analyze and describe images through a simple URL-based interface. It supports OpenAI’s vision models as well as OpenAI-compatible local models (such as LM Studio, Ollama, etc.), providing detailed descriptions of images and making it easy to integrate image analysis capabilities into your AI workflows.

Features

• Image Analysis: Analyze images from URLs and get detailed descriptions
• Flexible Model Support: Works with OpenAI’s vision models and OpenAI-compatible local models (LM Studio, Ollama, etc.)
• MCP Protocol: Fully compatible with the Model Context Protocol standard
• TypeScript: Built with TypeScript for type safety and better development experience
• Simple API: Easy-to-use interface for image description requests

Installation

Prerequisites

• Node.js 18+
• npm or yarn
• OpenAI API key or local vision model server (e.g., LM Studio, Ollama)

MCP Client Configuration

To use this server with an MCP client, add the following configuration:

{
  "mcpServers": {
    "image-recognition": {
      "command": "npx",
      "args": ["-y", "@akirose/image-recognition-mcp"],
      "env": {
        "OPENAI_API_KEY": "your-actual-openai-api-key-here"
      }
    }
  }
}

To allow access to image files from any path, set ALLOW_ALL_PATHS to true:

{
  "mcpServers": {
    "image-recognition": {
      "command": "npx",
      "args": ["-y", "@akirose/image-recognition-mcp"],
      "env": {
        "OPENAI_API_KEY": "your-actual-openai-api-key-here",
        "ALLOW_ALL_PATHS": "true"
      }
    }
  }
}

⚠️ IMPORTANT: The env section with your API key is required - this is the only way the MCP server can function. For local models, you can use any placeholder value for OPENAI_API_KEY and configure OPENAI_BASE_URL to point to your local server.

Environment Variables

The server supports the following environment variables:

• OPENAI_API_KEY - Your OpenAI API key, or any placeholder value when using local models (required)
• OPENAI_BASE_URL - Base URL for OpenAI API or OpenAI-compatible API servers (optional, defaults to OpenAI’s official API)
  • Example for LM Studio: "http://127.0.0.1:1234/v1"
  • Example for Ollama: "http://localhost:11434/v1"
• OPENAI_MODEL - The vision model to use for image recognition (optional, defaults to “gpt-5-mini”)
  • For OpenAI: "gpt-5-mini", "gpt-4o", "gpt-4o-mini", etc.
  • For local models: "llava", "qwen/qwen3-vl-4b", or any locally available vision model
• ALLOWED_IMAGE_PATHS - Comma-separated list of allowed local file paths (optional, defaults to “./images,./assets”)
  • Example: "./images,./assets,./downloads"