Happy slice of mcp pizzaMCP.pizza

aws-cost-explorer-mcp-server

MCP.Pizza Chef: aarora79

Four tools, all read-only: yesterday's compute spend, a day-by-day breakdown by region and service, and two views of your Amazon AI model usage pulled from logs. You sign in with the cloud credentials you already have rather than a key you sign up for, and read-only permission is enough. Two cautions worth knowing. The compute tool accepts a number-of-days setting and then ignores it, always returning one day. And Amazon bills you per request, so a chatty session costs real money.

Unmaintained · No commits in 17 months.
Data
Other

Use This MCP server To

See which service pushed my cloud bill up this month Break down yesterday's spending by region and service Check how much my AI model usage is costing Find the most expensive part of my cloud setup Pull spending figures for a second account I manage

README

AWS Cost Explorer and Amazon Bedrock Model Invocation Logs MCP Server & Client

An MCP server for getting AWS spend data via Cost Explorer and Amazon Bedrock usage data via Model invocation logs in Amazon Cloud Watch through Anthropic's MCP (Model Control Protocol). See section on "secure" remote MCP server to see how you can run your MCP server over HTTPS.

flowchart LR
    User([User]) --> UserApp[User Application]
    UserApp --> |Queries| Host[Host]
    
    subgraph "Claude Desktop"
        Host --> MCPClient[MCP Client]
    end
    
    MCPClient --> |MCP Protocol over HTTPS| MCPServer[AWS Cost Explorer MCP Server]
    
    subgraph "AWS Services"
        MCPServer --> |API Calls| CostExplorer[(AWS Cost Explorer)]
        MCPServer --> |API Calls| CloudWatchLogs[(AWS CloudWatch Logs)]
    end
Loading

You can run the MCP server locally and access it via the Claude Desktop or you could also run a Remote MCP server on Amazon EC2 and access it via a MCP client built into a LangGraph Agent.

🚨You can also use this MCP server to get AWS spend information from other accounts as long as the IAM role used by the MCP server can assume roles in those other accounts🚨

Demo video

AWS Cost Explorer MCP Server Deep Dive

Overview

This tool provides a convenient way to analyze and visualize AWS cloud spending data using Anthropic's Claude model as an interactive interface. It functions as an MCP server that exposes AWS Cost Explorer API functionality to Claude Desktop, allowing you to ask questions about your AWS spend in natural language.

Features

aws-cost-explorer-mcp-server FAQ

Do I need a key?
Not a sign-up key. You use the cloud credentials you already have, and read-only permission on the billing and logs services is enough.
Does using it cost anything?
Yes, a little. Amazon charges for each request to its cost data service, so an assistant asking many questions in a row adds real charges to your bill. Nothing here limits how often it asks.
Can it change anything in my cloud account?
No. Every tool only reads. It cannot start, stop, delete or reconfigure a thing.
Can I use this to work out why my bill went up?
Yes, and that is what it is best at. Ask for the day-by-day breakdown across services and regions, then keep asking follow-up questions in plain English.
Does the number-of-days setting actually work?
Not everywhere. The compute-spend tool accepts a days setting and quietly ignores it, always giving you a single day. The detailed breakdown does honour it.
How many tools does it add?
Four. Yesterday's compute spend, a detailed daily breakdown, and two views of AI model usage.
Do the AI usage tools work straight away?
No. They read logs that are switched off by default, so you have to turn on model invocation logging first, and storing those logs costs money of its own.
Can I look at a different account?
Yes, provided the identity you run it under is allowed to step into that account. You pass the account number along with your question.
Is it safe to run on a remote machine?
The documentation shows how to put it on a cloud server behind encryption, but there is no login of any kind. Anyone who can reach that machine can read your billing data.
How hard is setup?
This is a developer job. Expect Python, cloud credential setup, permission policies, and possibly a container.
Is it still being maintained?
It has been quiet since April 2025, well over a year. It is MIT licensed.