aws-finops

The AWS FinOps MCP Server exposes a set of tooling capabilities that connect your AWS financial data to your preferred chat or assistant workflow via MCP. It allows you to analyze AWS costs, surface waste, and produce budget insights using natural language prompts. The server runs locally and uses your existing AWS CLI configuration and IAM permissions, so credentials never leave your machine. In practice, you can ask questions like what was spent by a particular service last month, identify idle or unattached resources, or generate a summarized finops report across multiple profiles and regions. The server supports multi-profile and multi-region queries, enabling comprehensive visibility across your AWS environment from a single prompt. To use it with Claude for Desktop (or any MCP client), configure the MCP client to launch the AWS FinOps MCP Server, and then issue prompts that invoke the server's tools and return structured results suitable for your chat interface.

Prerequisites

• Python 3.10+
• MCP Client (examples include Claude for Desktop)
• AWS CLI installed and configured with at least one AWS profile
• An IAM policy with read-only permissions for cost and usage reporting (e.g., ce:GetCostAndUsage, budgets:ViewBudget, ec2:Describe*, sts:GetCallerIdentity)

Step 1: Install via pipx (recommended)

pipx install aws-finops-mcp-server

If you don't have pipx, install it and add to PATH:

python -m pip install --user pipx
python -m pipx ensurepath

Step 2: (Alternatively) Install via uv (for Python package installation) or from source Option 2: Using uv (Fast Python Package Installer)

# Install uv if you don't have it yet
curl -LsSf https://astral.sh/uv/install.sh | sh

# Create project directory
mkdir aws-finops-mcp-server && cd aws-finops-mcp-server

# Create and activate python virtual environment
uv venv && source .venv/bin/activate

# Install aws-finops-mcp-server
uv pip install aws-finops-mcp-server

Option 3: From Source

# Clone the repository
git clone https://github.com/ravikiranvm/aws-finops-mcp-server.git
cd aws-finops-mcp-server

# Create and activate python virtual environment
python -m venv .venv && source .venv/bin/activate

# Install using pip
pip install -e .

Step 3: AWS prerequisites

• Ensure AWS CLI is installed and configured with at least one profile: aws configure --profile your-profile-name
• Create an IAM policy with read-only permissions (the policy shown in the README) and attach it to your user or role

Usage note: After installation, you can configure MCP clients (e.g., Claude for Desktop) to run the server. For pipx installations, the executable is aws-finops-mcp-server and you can invoke it via the MCP client configuration using the provided command mapping.

Tips:

• Ensure your AWS CLI profiles are configured for each account/region you intend to query. The MCP server will leverage these profiles to access cost and usage data.
• The AWS FinOps MCP Server is designed to run locally; do not expose the server publicly. Keep your credentials and policies restricted to your machine.
• If you reorganize your AWS profiles or regions, update your MCP client configuration accordingly so prompts route to the correct data sources.
• When using Claude for Desktop, add the server entry to the mcpServers section of claude_desktop_config.json, using the name you chose (e.g., aws_finops).
• If you encounter PATH issues after installing via pipx, ensure the PATH includes the directory where the pipx shim is installed.