Skip to main content
Glama

vision-mcp

A visual MCP Server based on Alibaba Cloud Bailian DashScope (Qwen-VL). Provides image understanding capabilities for clients such as Claude Code, Trae, and CodeBuddy via the MCP protocol.

Features

  • analyze_image — General image analysis (supports local file paths / public URLs / data URLs)

  • screenshot_analyze — UI / interface screenshot analysis

Related MCP server: image-mcp

Supported Models

Model

Description

qwen-vl-max

Best quality, suitable for complex images

qwen-vl-plus

Balanced quality and speed (default)

qwen2.5-vl-72b-instruct

72B instruction following

qwen2.5-vl-7b-instruct

7B lightweight and fast

Quick Start

1. Install Dependencies

python -m venv .venv
# Windows
.venv\Scripts\pip install -r requirements.txt
# Linux / macOS
.venv/bin/pip install -r requirements.txt

2. Configure API Key

Create a .env file in the project root (automatically read when MCP Server starts):

DASHSCOPE_API_KEY=sk-你的密钥

Application URL: Alibaba Cloud Bailian Console

3. Start

# Windows
.venv\Scripts\python vision_mcp_server.py
# Linux / macOS
.venv/bin/python vision_mcp_server.py

List supported models:

python vision_mcp_server.py --list

Connect to MCP Client

Add the following to the client's MCP configuration (e.g., mcp.json or .mcp.json):

{
  "mcpServers": {
    "vision": {
      "command": "/path/to/vision-mcp/.venv/bin/python",
      "args": ["/path/to/vision-mcp/vision_mcp_server.py"],
      "env": {}
    }
  }
}

Testing

# 协议冒烟测试(无需 API Key)
.venv/bin/python smoke_test.py

# 真实调用测试(需配置 API Key,使用本地测试图)
.venv/bin/python live_test.py

Notes

  • The DashScope server may time out when downloading overseas public image URLs. It is recommended to use local file paths or URLs accessible within China

  • The API Key is provided via the .env file and is not written into the configuration file

License

MIT

Related MCP Connectors

Related MCP Servers

  • A
    license
    A
    quality
    A
    maintenance
    Multi-model vision understanding MCP server that provides unified image analysis for AI assistants without native vision, supporting models like GLM-4.6V, DeepSeek-OCR, Qwen3-VL-Flash, and more.
    1
    367 npm
    113
    MIT
  • A
    license
    Not graded
    quality
    C
    maintenance
    MCP server that adds Qwen3-VL vision capabilities to Claude Code or any MCP client, enabling OCR, UI/screenshot recognition, and chart understanding. It provides tools for analyzing single or batch images via DashScope's OpenAI-compatible API, with caching and fast mode.
    MIT
  • A
    license
    Not graded
    quality
    C
    maintenance
    MCP server that provides a 'borrowed eye' for text-only LLMs, enabling them to identify and describe local images via the Qwen VL vision model, including face recognition, scene description, OCR, and targeted visual questioning.
    3 npm
    Apache 2.0