vision-mcp
vision-mcp
基于 阿里云百炼 DashScope(Qwen-VL) 的视觉 MCP Server。通过 MCP 协议为 Claude Code、Trae、CodeBuddy 等客户端提供图片理解能力。
功能
analyze_image— 通用图片分析(支持本地文件路径 / 公网 URL / data URL)screenshot_analyze— UI / 界面截图分析
Related MCP server: image-mcp
支持的模型
模型 | 说明 |
| 效果最佳,适合复杂图片理解 |
| 效果与速度均衡(默认) |
| 72B 指令跟随 |
| 7B 轻量快速 |
快速开始
1. 安装依赖
python -m venv .venv
# Windows
.venv\Scripts\pip install -r requirements.txt
# Linux / macOS
.venv/bin/pip install -r requirements.txt2. 配置 API Key
在项目根目录创建 .env 文件(MCP Server 启动时自动读取):
DASHSCOPE_API_KEY=sk-你的密钥申请地址:阿里云百炼控制台
3. 启动
# Windows
.venv\Scripts\python vision_mcp_server.py
# Linux / macOS
.venv/bin/python vision_mcp_server.py列出支持的模型:
python vision_mcp_server.py --list接入 MCP 客户端
在客户端的 MCP 配置(如 mcp.json 或 .mcp.json)中添加:
{
"mcpServers": {
"vision": {
"command": "/path/to/vision-mcp/.venv/bin/python",
"args": ["/path/to/vision-mcp/vision_mcp_server.py"],
"env": {}
}
}
}测试
# 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注意
DashScope 服务端下载海外公网图片 URL 可能超时,建议使用本地文件路径或国内可达的 URL
API Key 通过
.env文件提供,不会写入配置文件
License
MIT
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
MCP server for Qwen Image 3 AI image generation
MCP server for ByteDance Seedream AI image generation
BeautySpark MCP server for ChatGPT, Claude and other MCP clients
MCP server for OpenAI API (chat completions, image generation, embeddings) via AceDataCloud
Related MCP Servers
- AlicenseAqualityCmaintenanceMulti-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.1512 npm116MIT
- AlicenseAqualityDmaintenanceMCP server for image recognition, supporting multiple vision backends (Anthropic, Zhipu, Ollama) to describe, answer questions, and analyze images.355 npm1MIT
- AlicenseNot gradedqualityCmaintenanceMCP 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
- FlicenseNot gradedqualityDmaintenanceMCP 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.-