aws-finops-mcp-server and aws-mcp-server
These are complements: the FinOps-focused tool provides cost analysis and optimization insights, while the general AWS CLI execution tool supplies the underlying infrastructure access needed to implement those recommendations and manage AWS resources.
About aws-finops-mcp-server
ravikiranvm/aws-finops-mcp-server
An MCP (Model Context Protocol) server that brings powerful AWS FinOps capabilities directly into your AI assistant. Analyze cloud costs, audit for waste, and get budget insights using natural language, all while keeping your credentials secure on your local machine.
Exposes AWS Cost Explorer and Budget APIs through two primary tools—`get_cost` for granular cost queries across services/tags/time ranges, and `run_finops_audit` for detecting infrastructure waste (stopped instances, unattached volumes, orphaned IPs)—while supporting multi-profile and multi-region queries via existing AWS CLI credentials. Implements the MCP stdio transport pattern, integrating natively with Claude Desktop and Amazon Q CLI as MCP clients, requiring only read-only IAM permissions (ce:GetCostAndUsage, budgets:ViewBudget, ec2:Describe*) with all credential handling delegated to local AWS CLI profiles.
About aws-mcp-server
alexei-led/aws-mcp-server
A lightweight service that enables AI assistants to execute AWS CLI commands (in safe containerized environment) through the Model Context Protocol (MCP). Bridges Claude, Cursor, and other MCP-aware AI tools with AWS CLI for enhanced cloud infrastructure management.
Exposes AWS CLI through two dynamically-documented tools (`aws_cli_help` and `aws_cli_pipeline`) that let Claude discover and execute commands on-demand rather than wrapping individual APIs. Implements multiple transport protocols (stdio, streamable-http) with configurable sandboxing modes and strict output/timeout limits, relying on IAM policies as the primary security boundary. Supports Docker containerization for host isolation and credential passing via environment variables or AWS config files, with graceful shutdown and proper MCP error handling.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work