About s3
S3 is an MCP server for AWS S3 that handles bucket and object operations with configurable tool groups. It suits users who need to limit what an AI client can do, and it also works with S3-compatible services such as MinIO and LocalStack.
Key features
- 01Bucket and object list, get, put, copy, and delete
- 02Tool groups separating read-only and read-write access
- 03Individual tool enable or disable through environment variables
- 04AWS credential validation with health checks
- 05S3-compatible endpoint and path-style addressing support
- 06Single-bucket mode that scopes all operations
Use cases
- 01Letting an AI assistant browse S3 contents in read-only mode
- 02Testing object storage workflows against MinIO or LocalStack
- 03Restricting AI access to one bucket
MCP support
This project is an MCP server that exposes S3 operations as tools to MCP clients, and provides an s3://config resource for viewing server configuration and status.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
npx -y s3-aws-mcp-serverAWS_ACCESS_KEY_IDSecretAWS access key ID for authentication.AWS_SECRET_ACCESS_KEYSecretAWS secret access key for authentication.AWS_REGIONAWS region for S3 operations.AWS_ENDPOINT_URLCustom S3 endpoint URL for MinIO or S3-compatible services.S3_FORCE_PATH_STYLEUse path-style addressing instead of virtual-hosted-style.S3_BUCKETConstrain all operations to a single bucket.S3_ENABLED_TOOLGROUPSComma-separated tool groups: readonly or readwrite.S3_ENABLED_TOOLSWhitelist specific tools by name.S3_DISABLED_TOOLSBlacklist specific tools by name.SKIP_HEALTH_CHECKSSkip AWS credential validation health checks at startup.Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMES3 MCP Server
MCP server for AWS S3 operations with fine-grained tool access control. Supports read/write operations on S3 buckets and objects, with configurable tool groups for security and permission management.
Highlights
- Full S3 bucket and object management (list, get, put, copy, delete)
- Fine-grained access control with tool groups (readonly, readwrite)
- Individual tool enable/disable via environment variables
- AWS credential validation with health checks
- S3-compatible endpoint support (MinIO, LocalStack, etc.)
- TypeScript with strict type checking
- Comprehensive testing (functional, integration, manual)
Capabilities
Tools
| Tool | Group | Description |
|---|---|---|
list_buckets | readonly | List all S3 buckets in the AWS account |
list_objects | readonly | List objects in a bucket with prefix and pagination |
get_object | readonly | Get object contents as text |
head_bucket | readonly | Check if a bucket exists and is accessible |
put_object | readwrite | Upload or update an object |
delete_object | readwrite | Delete an object from a bucket |
copy_object | readwrite | Copy an object within or across buckets |
create_bucket | readwrite | Create a new S3 bucket |
delete_bucket | readwrite | Delete an empty S3 bucket |
Resources
| Resource | Description |
|---|---|
s3://config | Server configuration and status (for debugging) |
Tool Groups
Control which tools are available via the S3_ENABLED_TOOLGROUPS environment variable:
| Group | Description |
|---|---|
readonly | Read-only operations (list, get, head) |
readwrite | Write operations (put, delete, copy, create) |
Examples:
S3_ENABLED_TOOLGROUPS="readonly"- Only read operationsS3_ENABLED_TOOLGROUPS="readonly,readwrite"- All operations- Not set - All tools enabled (default)
Individual Tool Control
Fine-grained control over specific tools:
S3_ENABLED_TOOLS="list_buckets,get_object"- Only enable these toolsS3_DISABLED_TOOLS="delete_bucket,delete_object"- Disable these tools
Single Bucket Mode
Constrain all operations to a specific bucket using S3_BUCKET:
S3_BUCKET="my-bucket"
When set:
- All object operations are automatically scoped to this bucket
- Bucket-level tools (
list_buckets,create_bucket,delete_bucket,head_bucket) are hidden - The
bucketparameter is automatically injected and hidden from tool inputs - For
copy_object, both source and destination are constrained to the specified bucket
This is useful for restricting access to a single bucket without giving broader S3 permissions.
Quick Start
Configuration
Environment Variables
| Variable | Required | Description | Default |
|---|---|---|---|
AWS_ACCESS_KEY_ID | Yes | AWS access key ID | - |
AWS_SECRET_ACCESS_KEY | Yes | AWS secret access key | - |
AWS_REGION | No | AWS region for S3 operations | us-east-1 |
AWS_ENDPOINT_URL | No | Custom S3 endpoint (for MinIO, etc.) | - |
S3_FORCE_PATH_STYLE | No | Use path-style addressing (for MinIO) | false |
S3_BUCKET | No | Constrain operations to single bucket | - |
S3_ENABLED_TOOLGROUPS | No | Comma-separated tool groups | All enabled |
S3_ENABLED_TOOLS | No | Specific tools to enable | - |
S3_DISABLED_TOOLS | No | Specific tools to disable | - |
SKIP_HEALTH_CHECKS | No | Skip credential validation | false |
Claude Desktop Configuration
If this is your first time using MCP Servers, make sure you have the Claude Desktop application and follow the official MCP setup instructions.
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"s3": {
"command": "npx",
"args": ["-y", "s3-aws-mcp-server"],
"env": {
"AWS_ACCESS_KEY_ID": "your-access-key-id",
"AWS_SECRET_ACCESS_KEY": "your-secret-access-key",
"AWS_REGION": "us-east-1",
"S3_ENABLED_TOOLGROUPS": "readonly"
}
}
}
}
Restart Claude Desktop and you should be ready to go!
Read-Only Mode
For safer exploration, enable only read operations:
{
"env": {
"S3_ENABLED_TOOLGROUPS": "readonly"
}
}
Using with S3-Compatible Services
For MinIO, LocalStack, or other S3-compatible services:
{
"env": {
"AWS_ACCESS_KEY_ID": "minioadmin",
"AWS_SECRET_ACCESS_KEY": "minioadmin",
"AWS_REGION": "us-east-1",
"AWS_ENDPOINT_URL": "http://localhost:9000",
"S3_FORCE_PATH_STYLE": "true"
}
}
Note: S3_FORCE_PATH_STYLE is required for MinIO and some other S3-compatible services that use path-style addressing (localhost:9000/bucket-name) instead of virtual-hosted-style (bucket-name.localhost:9000).
Development
Install Dependencies
npm run install-all
Build
npm run build
Running in Development Mode
npm run dev
Testing
# Run functional tests
npm run test:run
# Run integration tests (full MCP protocol)
npm run test:integration
# Run manual tests (real AWS - requires .env)
npm run test:manual:setup # First time only
npm run test:manual
# Run all automated tests
npm run test:all
Manual Testing Setup
Create a .env file in the s3 directory:
AWS_ACCESS_KEY_ID=your_access_key
AWS_SECRET_ACCESS_KEY=your_secret_key
AWS_REGION=us-east-1
Then run:
npm run test:manual
Project Structure
s3/
├── local/ # Local server implementation
│ ├── src/
│ │ ├── index.ts # Main entry point with env validation
│ │ └── index.integration-with-mock.ts
│ └── package.json
├── shared/ # Shared business logic
│ ├── src/
│ │ ├── server.ts # Server factory with DI
│ │ ├── tools.ts # Tool registration with grouping
│ │ ├── tools/ # Individual tool implementations
│ │ ├── resources.ts # Resource implementations
│ │ ├── s3-client/ # AWS S3 client wrapper
│ │ └── logging.ts
│ └── package.json
├── tests/ # Test suite
│ ├── functional/ # Unit tests with mocks
│ ├── integration/ # MCP protocol tests
│ ├── manual/ # Real API tests
│ └── mocks/ # Mock implementations
└── package.json # Root workspace config
License
MIT
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.