S3 MCP Server
Enables interaction with AWS S3 through MCP, supporting bucket and object management, lifecycle configurations, tagging, policies, CORS settings, presigned URLs, and file uploads/downloads.
README
Overview
This server enables you to:
- Manage S3 buckets and objects
- Handle lifecycle configurations
- Set and retrieve object tags
- Manage bucket policies
- Configure CORS settings
Installation
There are multiple ways to use this server depending on your setup.
Cursor (recommended)
Add this to your Cursor MCP configuration:
{
"mcpServers": {
"s3-mcp": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"AWS_SECRET_ACCESS_KEY=your_access_key",
"-e",
"AWS_ACCESS_KEY_ID=your_access_key",
"-e",
"AWS_REGION=your_region",
"ashgw/s3-mcp:latest"
]
}
}
}
If you prefer pinning to a specific Docker image build (e.g., 20250413-165732), use that tag instead of
latest. Browse available versions on Docker Hub.
Cursor will route that request through the MCP server automatically.
Make sure it's green and all the tools are available. <img width="951" height="430" alt="image" src="https://github.com/user-attachments/assets/c1c0b425-d44b-4cd7-b34d-1621bab2a05d" />
Docker (manual)
You can run the S3 MCP server manually via Docker:
docker run --rm -it \
-e AWS_ACCESS_KEY_ID=your_access_key \
-e AWS_SECRET_ACCESS_KEY=your_secret_key \
-e AWS_REGION=your_region \
ashgw/s3-mcp
This uses the pre-built image published at ashgw/s3-mcp.
Repo
Clone the repository and cd into it, then build with:
docker build -t s3-mcp .
Then run with:
docker run --rm -e AWS_ACCESS_KEY_ID=your_access_key -e AWS_SECRET_ACCESS_KEY=your_secret_key -e AWS_REGION=your_region s3-mcp
Environment Variables
Set the following environment variables for AWS credentials:
export AWS_ACCESS_KEY_ID=your_access_key
export AWS_SECRET_ACCESS_KEY=your_secret_key
export AWS_REGION=your_region
Features
- List Buckets: Retrieve a list of all S3 buckets.
- Create Bucket: Create a new S3 bucket with optional configurations.
- List Objects: List all objects in a specified bucket.
- Get Object: Retrieve the content of a specified object.
- Put Object: Upload an object to a specified bucket.
- Delete Object: Remove an object from a specified bucket.
- Generate Presigned URL: Create a presigned URL for accessing or uploading an object.
- Set Bucket Policy: Update or set a policy for a specified bucket.
- Get Bucket Policy: Retrieve the current policy for a specified bucket.
- Delete Bucket Policy: Remove the current policy for a specified bucket.
- Lifecycle Configuration: Manage lifecycle rules for S3 buckets.
- Object Tagging: Set and retrieve tags for S3 objects.
- CORS Configuration: Get and set CORS rules for a bucket.
- Copy Object: Copy an object from one location to another within S3.
- Download File to Local: Download a file from a specified S3 bucket to a local path.
- Upload Local File: Upload a local file to a specified S3 bucket.
Few Usage Examples
List Buckets
response = await tool("list_buckets")
print(response)
Create a Bucket
response = await tool("create_bucket", {
"bucket_name": "my-new-bucket",
"region": "us-west-1",
"config": {
"blockPublicAccess": {
"BlockPublicAcls": True,
"IgnorePublicAcls": True,
"BlockPublicPolicy": True,
"RestrictPublicBuckets": True
},
"versioning": True,
"encryption": "AES256"
}
})
print(response)
Upload a File
response = await tool("upload_local_file", {
"bucket_name": "my-new-bucket",
"local_path": "/path/to/local/file.txt",
"key": "file.txt"
})
print(response)
Get Object Tags
response = await tool("get_object_tagging", {
"bucket_name": "my-new-bucket",
"key": "file.txt"
})
print(response)
License
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。