The AWS MCP Server enables users to manage and query AWS resources through natural language commands via an AI assistant like Claude, executing AWS SDK V2 JavaScript code locally.
Key capabilities:
Natural Language AWS Operations: Query and manage EC2 instances, S3 buckets, Lambda functions, ECS clusters, and other AWS resources using conversational commands
Custom Code Execution: Run JavaScript code with AWS SDK V2 following best practices for optimization, error handling, pagination, and minimal data return
Credential Management: List, view, and switch between AWS profiles and credentials configured locally, with support for SSO authentication
Multi-Region Support: Interact with AWS resources across different regions
Secure Operations: Handle AWS credentials locally without external exposure
Claude Desktop Integration: Seamlessly connect with Claude Desktop for AI-driven AWS management
Integrates with AWS (Amazon Web Services), allowing Claude to query and manage AWS resources through natural language. Supports listing EC2 instances, S3 buckets, Lambda functions, ECS clusters, and other AWS resources across multiple regions and profiles.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@AWS MCP Server list all EC2 instances in my account"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
MCP (Model Context Protocol) is an open protocol that standardizes how applications provide context to large language models (LLMs). Think of MCP like a USB-C port for AI applications. Just as USB-C provides a standard method for connecting devices to various accessories, MCP provides a standard method for connecting AI models to different data sources and tools.
MCP enables building complex agents and automated workflows based on LLMs. LLMs often need to integrate with data and tools, and MCP provides:
A growing list of ready-to-use integrations that allow your LLM to connect directly.
Flexibility to switch between LLM providers.
Best data security practices within your infrastructure.
At its core, MCP follows a client-server architecture where a host application can connect to multiple servers:
MCP Hosts: Programs like Claude Desktop, IDEs, or AI tools that want to access data through MCP.
MCP Clients: Protocol clients that maintain 1:1 connections with servers.
MCP Servers: Lightweight programs, each providing specific capabilities through the Model Context Protocol.
Local Data Sources: Files, databases, and services on your computer that MCP servers can securely access.
Remote Services: External systems available on the Internet (e.g., via APIs) that MCP servers can connect to.
Related MCP server: AWS Resources MCP Server
AWS MCP (Model Context Protocol) is a server that enables AI assistants like Claude to interact with AWS environments through natural language. This makes it easy to manage and query AWS resources without using the traditional AWS Console or CLI.
AWS MCP can be considered a powerful alternative to Amazon Q, offering greater flexibility and security.
Query and modify AWS resources using natural language
Support for multiple AWS profiles and SSO authentication
Multi-region AWS support
Secure credential management (credentials are never exposed externally, only used locally)
Local execution with your AWS credentials
To use AWS MCP, ensure your environment has:
Node.js
Claude Desktop application
Locally configured AWS credentials (AWS Configure) (stored in ~/.aws/)
First, download the source code from GitHub to your computer:
git clone https://github.com/ihatesea69/AWS-MCP
cd aws-mcpYou can use pnpm or npm to install:
pnpm install
# or
npm installOpen the Claude Desktop application and navigate to:
Settings -> Developer -> Edit ConfigThen add the following entry to the claude_desktop_config.json file:
{
"mcpServers": {
"aws": {
"command": "npm", // or pnpm
"args": [
"--silent",
"--prefix",
"/Users/<YOUR USERNAME>/aws-mcp",
"start"
]
}
}
}Note: Replace
/Users/<YOUR USERNAME>/aws-mcpwith the actual path to your project directory.
Run your project by navigating (cd) to the directory containing index.ts and running the NPM start command.
After editing the configuration, restart the Claude Desktop application. If the installation is correct, you will see a successful connection message with MCP.
You can start using it by entering natural language commands such as:
"List available AWS profiles"
"List all EC2 instances in my account"
"Show me S3 buckets with their sizes"
"What Lambda functions are deployed in us-east-1?"
"List all ECS clusters and their services"
nvmIf you use Node.js through nvm, compile from source first and add the following configuration:
{
"mcpServers": {
"aws": {
"command": "/Users/<USERNAME>/.nvm/versions/node/v20.10.0/bin/node",
"args": [
"<WORKSPACE_PATH>/aws-mcp/node_modules/tsx/dist/cli.mjs",
"<WORKSPACE_PATH>/aws-mcp/index.ts",
"--prefix",
"<WORKSPACE_PATH>/aws-mcp",
"start"
]
}
}
}Note: Replace
<USERNAME>and<WORKSPACE_PATH>with the actual paths on your system.
AWS MCP is a powerful tool that allows you to manage AWS resources using the Claude AI assistant. With the ability to query and control through natural language, AWS MCP significantly simplifies AWS management. If you are looking for an alternative to Amazon Q, AWS MCP is a worthy consideration!
Original source: https://github.com/RafalWilinski/aws-mcp
List all AWS credentials/configs/profiles that are configured/usable on this machine
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden of behavioral disclosure. It implies a read-only operation by using 'List,' but doesn't specify if it requires permissions, how it handles errors, or what the output format looks like. For a tool with zero annotation coverage, this lack of detail on behavior is a notable shortfall.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single, efficient sentence that directly states the tool's function without any fluff. It is front-loaded with the core action and resource, making it easy to parse. Every word contributes to understanding the purpose, earning a top score for conciseness.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's simplicity (0 parameters, no output schema, no annotations), the description is adequate but has clear gaps. It explains what the tool does but lacks details on usage guidelines, behavioral traits, and output format. For a basic list tool, this is minimally viable but could be more complete by addressing when to use it and what to expect in return.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The tool has 0 parameters, and schema description coverage is 100%, so there are no parameters to document. The description doesn't need to add parameter semantics, but it does imply the scope ('on this machine'), which is useful context. Baseline for 0 parameters is 4, as the description adequately covers the tool's intent without unnecessary parameter details.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's purpose: 'List all AWS credentials/configs/profiles that are configured/usable on this machine.' It specifies the verb ('List') and resource ('AWS credentials/configs/profiles'), making the action explicit. However, it does not explicitly differentiate from sibling tools like 'select-profile' or 'run-aws-code', which could involve similar resources, so it doesn't reach a perfect score.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides no guidance on when to use this tool versus alternatives. It doesn't mention sibling tools or contexts where this tool is preferred, such as for inventory purposes before selecting a profile. Without any usage context, the agent must infer when to invoke it, which is a significant gap.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Run AWS code
| Name | Required | Description | Default |
|---|---|---|---|
| reasoning | Yes | The reasoning behind the code | |
| code | Yes | Your job is to answer questions about AWS environment by writing Javascript code using AWS SDK V2. The code must be adhering to a few rules: - Must be preferring promises over callbacks - Think step-by-step before writing the code, approach it logically - MUST written in Javascript (NodeJS) using AWS-SDK V2 - Avoid hardcoded values like ARNs - Code written should be as parallel as possible enabling the fastest and the most optimal execution - Code should be handling errors gracefully, especially when doing multiple SDK calls (e.g. when mapping over an array). Each error should be handled and logged with a reason, script should continue to run despite errors - DO NOT require or import "aws-sdk", it is already available as "AWS" variable - Access to 3rd party libraries apart from "aws-sdk" is not allowed or possible - Data returned from AWS-SDK must be returned as JSON containing only the minimal amount of data that is needed to answer the question. All extra data must be filtered out - Code MUST "return" a value: string, number, boolean or JSON object. If code does not return anything, it will be considered as FAILED - Whenever tool/function call fails, retry it 3 times before giving up with an improved version of the code based on the returned feedback - When listing resources, ensure pagination is handled correctly so that all resources are returned - Do not include any comments in the code - When doing reduce, don't forget to provide an initial value - Try to write code that returns as few data as possible to answer without any additional processing required after the code is run - This tool can ONLY write code that interacts with AWS. It CANNOT generate charts, tables, graphs, etc. Please use artifacts for that instead Be concise, professional and to the point. Do not give generic advice, always reply with detailed & contextual data sourced from the current AWS environment. Assume user always wants to proceed, do not ask for confirmation. I'll tip you $200 if you do this right. | |
| profileName | No | Name of the AWS profile to use | |
| region | No | Region to use (if not provided, us-east-1 is used) |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden of behavioral disclosure but offers minimal information. 'Run AWS code' suggests execution but doesn't disclose whether this is a read or write operation, what permissions are required, whether it has side effects, rate limits, or error handling behavior. The description fails to provide the behavioral context needed for safe and effective tool invocation.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is extremely concise at just two words, with zero wasted language. While this conciseness comes at the expense of completeness, from a pure structural perspective, it's front-loaded and contains no unnecessary verbiage. Every word in the description directly relates to the tool's function.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the complexity of running AWS code (which involves execution, potential mutations, security implications, and error handling), the description is completely inadequate. With no annotations, no output schema, and a minimal description, an agent lacks critical information about what this tool actually does, what it returns, and how to use it safely. The description fails to provide the contextual completeness needed for a tool of this nature.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
With 100% schema description coverage, the input schema already documents all four parameters thoroughly. The description adds no additional parameter semantics beyond what's already in the schema. The baseline score of 3 reflects that the schema does the heavy lifting, though the description could have provided higher-level context about how parameters relate to each other or typical usage patterns.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description 'Run AWS code' is essentially a tautology that restates the tool name without providing meaningful specificity. It doesn't clarify what type of AWS code is run, what resources it interacts with, or what distinguishes it from sibling tools like list-credentials and select-profile. The description lacks a clear verb+resource combination that would help an agent understand the tool's actual function.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides absolutely no guidance about when to use this tool versus alternatives. There's no mention of appropriate contexts, prerequisites, or relationships to sibling tools. An agent would have no information about whether this tool should be used for AWS operations versus the other available tools, making selection decisions difficult.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Selects AWS profile to use for subsequent interactions. If needed, does SSO authentication
| Name | Required | Description | Default |
|---|---|---|---|
| profile | Yes | Name of the AWS profile to select | |
| region | No | Region to use (if not provided, us-east-1 is used) |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden of behavioral disclosure. It adds useful context about SSO authentication ('If needed, does SSO authentication'), which is not inferable from the input schema. However, it lacks details on side effects (e.g., whether this persists across sessions), error handling, or what 'subsequent interactions' specifically entails, leaving gaps in behavioral understanding.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is extremely concise (two short sentences) and front-loaded with the primary purpose. Every word earns its place, with no redundant or vague phrasing, making it efficient and easy to parse.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's moderate complexity (configures context for AWS operations) and lack of annotations or output schema, the description is mostly complete. It covers the purpose, conditional SSO behavior, and usage context. However, it misses details on return values or error cases, which could be important for a tool that sets state, slightly reducing completeness.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The schema description coverage is 100%, so the schema already fully documents both parameters ('profile' and 'region'). The description does not add any meaning beyond what the schema provides (e.g., it doesn't explain profile naming conventions or region implications), resulting in the baseline score of 3 for adequate but no extra value.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the specific action ('Selects AWS profile') and resource ('for subsequent interactions'), with explicit mention of SSO authentication as a conditional behavior. It distinguishes from sibling tools like 'list-credentials' (which lists rather than selects) and 'run-aws-code' (which executes code rather than configuring context).
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides clear context for when to use this tool ('for subsequent interactions') and implies it should be used before other AWS operations. However, it does not explicitly state when not to use it or name alternatives (e.g., when to use 'list-credentials' instead), which prevents a perfect score.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
list-credentialsrun-aws-codeselect-profileScored across 3 tools
Each tool has a clearly distinct purpose: list-credentials handles credential enumeration, run-aws-code executes AWS code, and select-profile manages profile selection and authentication. There is no overlap in functionality, making tool selection unambiguous for an agent.
The tools follow a consistent verb-noun pattern with hyphens (list-credentials, run-aws-code, select-profile), which is predictable and readable. The minor deviation is that 'aws-code' includes a hyphenated noun, but this does not break the overall naming convention.
With only 3 tools, the server feels thin for an AWS integration, as it lacks core operations like managing resources (e.g., EC2, S3) or performing common AWS tasks. However, the tools cover basic setup and execution, making it borderline appropriate for a minimal scope.
The tool surface is significantly incomplete for an AWS server, as it only handles credential management and code execution without any tools for interacting with AWS services (e.g., creating instances, uploading files, querying databases). This will likely cause agent failures when attempting broader AWS operations.
The AWS Knowledge MCP server is a fully managed remote Model Context Protocol server that provides real-time access to official AWS content in an LLM-compatible format. It offers structured access to AWS documentation, code samples, blog posts, What's New announcements, Well-Architected best practices, and regional availability information for AWS APIs and CloudFormation resources. Key capabilities include searching and reading documentation in markdown format, getting content recommendations, listing AWS regions, and checking regional availability for services and features.
Hosted Amazon Seller Central and Amazon Ads MCP server for Claude, ChatGPT, Cursor, and agents.
Hosted Amazon Seller and Vendor MCP server for Claude, ChatGPT, Cursor, Codex, Gemini, Copilot.
Hosted MCP server connecting claude.ai, ChatGPT and other AI apps to your own computer
