mcp-server-mimo-vision

MCP server for image vision analysis using multimodal models (MiMo, Claude, GPT-4o, etc.).

Why?

Claude Code's main model may not support multimodal (vision) input. This MCP server bridges the gap by calling a vision-capable model to analyze images and return text descriptions.

Quick Start

Option 1: npx (Recommended)

Add to .mcp.json in your project or ~/.claude/.mcp.json:

{
  "mcpServers": {
    "vision": {
      "command": "npx",
      "args": ["-y", "@ruqingwang/mcp-server-mimo-vision"],
      "env": {
        "ANTHROPIC_AUTH_TOKEN": "your-token-here"
      }
    }
  }
}

Option 2: One-click Install Script

curl -fsSL https://raw.githubusercontent.com/wangruqing723/mimo-vision-mcp/main/install.sh | bash

Option 3: Global Install

npm install -g @ruqingwang/mcp-server-mimo-vision

Configuration

After install, all settings are in ~/.claude/vision-mcp.conf:

ANTHROPIC_BASE_URL="https://your-proxy.com/anthropic"
ANTHROPIC_AUTH_TOKEN="your-token"
VISION_MODEL="mimo-v2.5"  # Change to any vision model

Supported Models

Any model compatible with Anthropic Messages API that supports multimodal:

| Model | Description | |-------|-------------| | mimo-v2.5 | MiMo multimodal (default) | | mimo-v2.5-pro | MiMo Pro | | claude-sonnet-4-20250514 | Claude Sonnet | | gpt-4o | GPT-4o (requires compatible proxy) |

Environment Variables (override config file)

| Variable | Description | Default | |----------|-------------|---------| | ANTHROPIC_BASE_URL | API proxy URL | https://token-plan-cn.xiaomimimo.com/anthropic | | ANTHROPIC_AUTH_TOKEN | API auth token | Required | | VISION_MODEL | Model name | mimo-v2.5 |

Image Input Modes

The describe_image and ocr_image tools accept:

1. File path: /path/to/image.png
2. HTTP URL: https://example.com/img.jpg
3. Data URI: data:image/png;base64,...
4. Raw base64: base64-encoded image string

Auto-intercept Hook (Optional)

To automatically use vision MCP when Claude Code reads an image file, add a PreToolUse hook to ~/.claude/settings.json:

{
  "hooks": {
    "PreToolUse": [
      {
        "matcher": "Read",
        "hooks": [
          {
            "type": "command",
            "command": "/path/to/hook-vision.sh",
            "statusMessage": "Analyzing image..."
          }
        ]
      }
    ]
  }
}

Tools

| Tool | Description | |------|-------------| | describe_image | Analyze image content, return text description | | ocr_image | Extract text from image (OCR) |

Usage

Once configured, just ask in Claude Code:

Look at /path/to/image.png and describe what's in it

Extract text from /path/to/screenshot.png

GitHub

https://github.com/wangruqing723/mimo-vision-mcp

License

MIT