MCP Server for Image & Video Analysis

中文文档 | English

English

A customizable Model Context Protocol (MCP) server that provides AI capabilities for image and video analysis with support for multiple AI providers and self-hosted APIs.

🌟 Features

• 🔧 Fully Customizable - Support for multiple AI providers and self-hosted APIs
• 🎯 Multiple AI Tools - UI generation, text extraction, error diagnosis, diagram analysis, and more
• 🚀 High Performance - Optimized for vision and multimodal tasks
• 🔄 Flexible Configuration - Easy to adapt to different AI platforms
• 📦 Ready to Use - Pre-compiled for quick deployment

📦 Installation

# Install from npm
npm install chuangjian-jianyanzhongxin/mcp-server

# Or use yarn
yarn add chuangjian-jianyanzhongxin/mcp-server

🚀 Quick Start

# 1. Set up environment variables
export ANTHROPIC_BASE_URL=http://127.0.0.1:8000
export ANTHROPIC_API_KEY=your-api-key-here
export ANTHROPIC_MODEL=your-model-name

# 2. Add the MCP server to your client configuration
# Claude Code:
claude mcp add mcp-server \
  --env ANTHROPIC_BASE_URL=http://127.0.0.1:8000 \
  --env ANTHROPIC_API_KEY=your-api-key \
  --env ANTHROPIC_MODEL=your-model-name \
  -- npx chuangjian-jianyanzhongxin/mcp-server

# Or manually add to your client's settings.json:
# {
#   "mcpServers": {
#     "mcp-server": {
#       "command": "node",
#       "args": ["node_modules/chuangjian-jianyanzhongxin/mcp-server/build/index.js"],
#       "env": {
#         "ANTHROPIC_BASE_URL": "http://127.0.0.1:8000",
#         "ANTHROPIC_API_KEY": "your-api-key",
#         "ANTHROPIC_MODEL": "your-model-name"
#       }
#     }
#   }
# }

⚙️ Configuration

Required Environment Variables

| Variable | Description | Example | |----------|-------------|---------| | ANTHROPIC_BASE_URL | Your API server URL | http://127.0.0.1:8000 | | ANTHROPIC_API_KEY | Your API key | your-api-key-here |

Optional Environment Variables

| Variable | Description | Default | |----------|-------------|---------| | ANTHROPIC_MODEL | Model name | Auto-detected | | ANTHROPIC_MODEL_TEMPERATURE | Temperature for generation | 0.8 | | ANTHROPIC_MODEL_TOP_P | Top P parameter | 0.6 | | ANTHROPIC_MODEL_MAX_TOKENS | Maximum tokens | 32768 | | ANTHROPIC_TIMEOUT | Request timeout (ms) | 300000 | | ANTHROPIC_RETRY_COUNT | Retry count on failure | 1 |

🛠️ Available Tools

1. ui_to_artifact - Convert UI screenshots to code, prompts, specs, or descriptions
2. extract_text_from_screenshot - OCR and text extraction from images
3. diagnose_error_screenshot - Error diagnosis and troubleshooting
4. understand_technical_diagram - Technical diagram analysis
5. analyze_data_visualization - Data visualization insights
6. ui_diff_check - UI comparison for visual regression
7. analyze_image - General-purpose image analysis
8. analyze_video - Video content analysis

📋 Requirements

• Node.js 18.0.0 or higher
• npm or yarn package manager
• MCP client (Claude Code, etc.)

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

📄 License

MIT License - see LICENSE file for details.

📚 Documentation

• Configuration Guide
• API Documentation - Coming soon
• Troubleshooting Guide - Coming soon

中文文档

一个可自定义的 Model Context Protocol (MCP) 服务器，提供 AI 图像和视频分析能力，支持多种 AI 提供商和自托管 API。

✨ 功能特点

• 🔧 高度可配置 - 支持多种 AI 提供商和自托管 API
• 🎯 多种 AI 工具 - UI 生成、文本提取、错误诊断、图表分析等
• 🚀 高性能 - 针对 vision 和 multimodal 任务进行优化
• 🔄 灵活配置 - 易于适配不同的 AI 平台
• 📦 开箱即用 - 预编译，快速部署

📦 安装

# 从 npm 安装
npm install chuangjian-jianyanzhongxin/mcp-server

# 或使用 yarn
yarn add chuangjian-jianyanzhongxin/mcp-server

🚀 快速开始

# 1. 设置环境变量
export ANTHROPIC_BASE_URL=http://127.0.0.1:8000
export ANTHROPIC_API_KEY=your-api-key-here
export ANTHROPIC_MODEL=your-model-name

# 2. 将 MCP 服务器添加到你的客户端配置
# Claude Code:
claude mcp add mcp-server \
  --env ANTHROPIC_BASE_URL=http://127.0.0.1:8000 \
  --env ANTHROPIC_API_KEY=your-api-key \
  --env ANTHROPIC_MODEL=your-model-name \
  -- npx chuangjian-jianyanzhongxin/mcp-server

# 或手动添加到客户端的 settings.json:
# {
#   "mcpServers": {
#     "mcp-server": {
#       "command": "node",
#       "args": ["node_modules/chuangjian-jianyanzhongxin/mcp-server/build/index.js"],
#       "env": {
#         "ANTHROPIC_BASE_URL": "http://127.0.0.1:8000",
#         "ANTHROPIC_API_KEY": "your-api-key",
#         "ANTHROPIC_MODEL": "your-model-name"
#       }
#     }
#   }
# }

⚙️ 配置

必需的环境变量

| 变量名 | 说明 | 示例 | |--------|------|------| | ANTHROPIC_BASE_URL | 你的 API 服务器地址 | http://127.0.0.1:8000 | | ANTHROPIC_API_KEY | 你的 API 密钥 | your-api-key-here |

可选的环境变量

| 变量名 | 说明 | 默认值 | |--------|------|--------| | ANTHROPIC_MODEL | 模型名称 | 自动检测 | | ANTHROPIC_MODEL_TEMPERATURE | 生成温度 | 0.8 | | ANTHROPIC_MODEL_TOP_P | Top P 参数 | 0.6 | | ANTHROPIC_MODEL_MAX_TOKENS | 最大 token 数 | 32768 | | ANTHROPIC_TIMEOUT | 请求超时时间（毫秒） | 300000 | | ANTHROPIC_RETRY_COUNT | 失败重试次数 | 1 |

🛠️ 可用工具

1. ui_to_artifact - 将 UI 截图转换为代码、提示词、规范或描述
2. extract_text_from_screenshot - 从图片中提取文本和 OCR
3. diagnose_error_screenshot - 错误诊断和故障排查
4. understand_technical_diagram - 技术图表分析
5. analyze_data_visualization - 数据可视化洞察
6. ui_diff_check - UI 对比和视觉回归检测
7. analyze_image - 通用图像分析
8. analyze_video - 视频内容分析

📋 系统要求

• Node.js 18.0.0 或更高版本
• npm 或 yarn 包管理器
• MCP 客户端（Claude Code 等）

🤝 贡献

欢迎贡献！请随时提交 Pull Request。

📄 许可证

MIT 许可证 - 详见 LICENSE 文件。

📚 文档

• 配置指南
• API 文档 - 即将推出
• 故障排除指南 - 即将推出