AWS Cost MCP Server

An MCP server focused on querying AWS Cost and Usage Report (CUR) data through Amazon Athena. This server provides tools to analyze AWS cost data efficiently through the Model Context Protocol.

Features

• Query AWS Cost and Usage Report data through Athena
  • More cost-effective than Cost Explorer API (pay only for the queries you run)
  • Access to detailed, raw cost data with granular resource-level insights
  • Access to historical data to do in-depth cost trend analysis
  • No API rate limits or throttling constraints
• Execute and manage Athena queries

Prerequisites

• Python 3.10+
• AWS credentials configured with access to Athena and CUR
• AWS Cost and Usage Reports set up with Athena integration

Installation

1. Clone the repository:

   git clone https://github.com/CRED-CLUB/aws-cost-mcp.git
   cd aws-cost-mcp
   

2. Run the setup script:

   chmod +x setup.sh
   ./setup.sh
   

   The setup script will:

   • Create a Python virtual environment
   • Install required dependencies
   • Create a .env file from .env.example

3. Configure your environment:

   Edit the .env file with your settings:

   # MCP Server Configuration
   MCP_SERVER_HOST=localhost
   MCP_SERVER_PORT=5004
   MCP_LOG_LEVEL=INFO
   
   # AWS Configuration
   AWS_REGION=us-east-1
   AWS_ACCESS_KEY_ID=your_access_key
   AWS_SECRET_ACCESS_KEY=your_secret_key
   
   # Athena Configuration (Optional - used as defaults if not provided in queries)
   ATHENA_OUTPUT_LOCATION=s3://your-bucket/athena-results/
   ATHENA_DATABASE=your_cur_database
   ATHENA_TABLE=your_cur_table
   ATHENA_WORKGROUP=primary
   ATHENA_CATALOG=your_cur_catalog(AwsDataCatalog)
   

Connecting to Cline

To connect this MCP server to Cline, you need to update the cline_mcp_settings.json file.

Add a configuration like the following to the mcpServers section:

{
  "mcpServers": {
    "aws": {
      "command": "absolute_path..../aws-cost-mcp/venv/bin/python",
      "args": [
        "absolute_path..../aws-cost-mcp/src/server.py"
      ],
      "env": {},
      "disabled": false,
      "autoApprove": [],
      "workingDirectory": "absolute_path..../aws-cost-mcp"
    }
  }
}

You can configure CLINE with the Custom Instructions available in cline_custom_instructions.txt to receive a detailed cost analysis.

Example query:

Using only the configured aws-cost-mcp server's resources and tools, analyze the ec2 cost trend across jan and feb 2025.

Cline response:

<img width="805" alt="Screenshot 2025-03-24 at 9 33 15 PM" src="https://github.com/user-attachments/assets/1808240d-8404-41a7-8455-598746a02070" />

Usage

Running the MCP Server

For Development and Testing

Use the MCP CLI's development mode, which provides interactive testing capabilities:

mcp dev src/server.py

For Production

Use the MCP CLI's run command:

mcp run src/server.py

By default, the server will use the configuration in the .env file:

MCP_SERVER_HOST=127.0.0.1
MCP_SERVER_PORT=8080
MCP_LOG_LEVEL=INFO

Documentation

The project documentation is organized in the cline_docs directory:

• projectRoadmap.md: Project goals, features, and progress tracking
• currentTask.md: Current objectives and next steps
• techStack.md: Technology choices and architecture decisions
• codebaseSummary.md: Project structure and component overview

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the terms of the LICENSE file.