Happy slice of mcp pizzaMCP.pizza

mcp-server-aws

MCP.Pizza Chef: rishikavikondala

Twenty-three tools cover Amazon's S3 storage and its DynamoDB database: listing buckets, reading and uploading files, creating tables, and adding, querying or removing records. It runs on your own machine and signs in with an Amazon Web Services access key belonging to an account that can both read and write. The deletion tools carry no confirmation step and there is no read-only mode, so a misread request can erase a bucket or a table. The project has had no code changes since April 2025.

Unmaintained · No commits in 17 months.
Data
Files/PDF

Use This MCP server To

List every storage bucket in my Amazon account Read a text file straight out of cloud storage Upload a file to a bucket without opening the console Look up a single record in one of my tables Search a table for rows matching a condition Create a new table for a small side project

README

AWS MCP Server

smithery badge

A Model Context Protocol server implementation for AWS operations that currently supports S3 and DynamoDB services. All operations are automatically logged and can be accessed through the audit://aws-operations resource endpoint.

AWS Server MCP server

See a demo video here.

Listed as a Community Server within the MCP servers repository.

Running locally with the Claude desktop app

Installing via Smithery

To install AWS MCP Server for Claude Desktop automatically via Smithery:

npx -y @smithery/cli install mcp-server-aws --client claude

Manual Installation

  1. Clone this repository.
  2. Set up your AWS credentials via one of the two methods below. Note that this server requires an IAM user with RW permissions for your AWS account for S3 and DynamoDB.
  • Environment variables: AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_REGION (defaults to us-east-1)
  • Default AWS credential chain (set up via AWS CLI with aws configure)
  1. Add the following to your claude_desktop_config.json file:
  • On MacOS: ~/Library/Application\ Support/Claude/claude_desktop_config.json
  • On Windows: %APPDATA%/Claude/claude_desktop_config.json
"mcpServers": {
  "mcp-server-aws": {
    "command": "uv",
    "args": [
      "--directory",
      "/path/to/repo/mcp-server-aws",
      "run",
      "mcp-server-aws"
    ]
  }
}
  1. Install and open the Claude desktop app.
  2. Try asking Claude to do a read/write operation of some sort to confirm the setup (e.g. create an S3 bucket and give it a random name). If there are issues, use the Debugging tools provided in the MCP documentation here.

mcp-server-aws FAQ

Is this project still being worked on?
It is not shut down, but the code has not changed since April 2025. Nothing about Amazon's storage or database service has broken it so far, though nobody appears to be watching.
Do I need an API key?
Not exactly. You need an Amazon Web Services access key and secret belonging to an account with permission to read and write S3 and DynamoDB. That is the key to your cloud account, and anything it does is billed to you.
Can it delete my files or tables?
Yes, and that is the main risk. Deleting a bucket, a file, a table or a record happens the moment it is asked, with no confirmation prompt and no read-only setting to fall back on.
Which apps does it work in?
Claude Desktop is the documented one, and any other app where you can paste a small local server configuration. It only runs on your own computer.
How many tools does it actually have?
Twenty-three. The list in the project's own notes leaves one out, a tool that lists your database tables, so you get slightly more than advertised.
Can I use this to create a new storage bucket?
Usually yes, but not in the default region. The code always attaches a region setting to the request, and Amazon rejects that for its us-east-1 region, so bucket creation fails there until you set a different region.
Does it keep a record of everything it does?
Partly. It offers an activity log, but reading a file's contents returns before the logging step and never shows up. The log also lives in memory and is wiped every time the server restarts.
How hard is setup?
Hard. You download the code yourself, need Python 3.13 or newer, and it is not published to any package index, so there is no one-line install.