本项目为 Claude 的 Model Context Protocol (MCP) 示例项目提供支持,展示了如何借助 Docker 和本地 Python 脚本扩展 Claude 的功能。项目提供了天气、TMDB 电影、Amity 大学信息以及文件系统访问四个服务器,助力用户丰富与 Claude 的交互体验。
编辑 claude_desktop_config.json 文件,添加如下配置:
{
"mcpServers": {
"weather": {
"command": "docker",
"args": [
"run",
"--rm",
"kunal/mcp-weather:latest"
]
},
"tmdb": {
"command": "docker",
"args": [
"run",
"--rm",
"kunal/mcp-tmdb:latest"
]
},
"amity-university": {
"command": "docker",
"args": [
"run",
"--rm",
"kunal/mcp-amity:latest"
]
},
"file-system": {
"command": "docker",
"args": [
"run",
"--rm",
"--mount",
"type=bind,src=/path/to/files,dst=/projects/files",
"kunal/mcp-file-system:latest",
"/projects"
]
}
}
}
编辑 claude_desktop_config.json 文件,添加以下配置:
{
"mcpServers": {
"weather": {
"command": "$(which uv)",
"args": [
"--directory",
"/path/to/mcp-weather",
"run",
"weather.py"
]
},
"tmdb": {
"command": "$(which uv)",
"args": [
"--directory",
"/path/to/mcp-tmdb",
"run",
"tmdb.py"
]
},
"amity-university": {
"command": "$(which uv)",
"args": [
"--directory",
"/path/to/mcp-amity",
"run",
"amity_crawler.py"
]
}
}
}
本项目通过 Docker 和本地 Python 脚本扩展了 Claude 的功能,提供了四个实用的服务器:
docker --version
docker pull ghcr.io/anthropic/uv-cli:latest
which uv # macOS/Linux
(Get-Command uv).Source # Windows
docker run hello-world 测试容器运行。docker images | grep kunal/mcp-。此文档为扩展 Claude 功能提供了详细指南。通过 Docker 和本地脚本的结合,您可以轻松添加自定义服务器,丰富交互体验。