MCP server/cloud

AWS FinOps MCP Server

Comprehensive AWS cloud resource optimization, cost management, and monitoring.

prashantgupta123/aws-pillar-mcp-server ↗by prashantgupta123updated
1

Add it to Claude Code

claude mcp add aws-finops-mcp -- python -m aws_finops_mcp
2

Make your agent remember this setup

aws-finops-mcp's config, env vars, and the gotchas you hit — recalled in every future Claude Code, Cursor, and Codex session.

npx conare@latest

Free · one command · indexes the sessions already on disk. Set up in the browser instead →

What it does

  • 76 tools across 14 categories for AWS resource management
  • Category-based tool filtering for 67-89% faster loading
  • Support for both stdio and HTTP modes
  • Comprehensive cost savings and security compliance analysis
  • Optimized for deployment on Amazon Bedrock AgentCore

Tools 1

find_unused_lambda_functionIdentifies unused AWS Lambda functions to reduce costs.

Environment Variables

MCP_TOOL_CATEGORIESComma-separated list of tool categories to enable for faster loading.

Try it

Find unused resources in us-east-1.
Analyze my AWS cost trends for the last month.
Identify unencrypted S3 buckets in my account.
List all Lambda functions that have not been invoked in the last 30 days.
Suggest right-sizing options for my over-provisioned EC2 instances.
Original README from prashantgupta123/aws-pillar-mcp-server

AWS FinOps MCP Server

Production-Ready Deployment: This MCP server is optimized for deployment on Amazon Bedrock AgentCore Runtime

Model Context Protocol (MCP) server for AWS Financial Operations (FinOps) - providing comprehensive tools for cloud resource optimization, cost management, and performance monitoring.

🚀 Quick Start - Deploy to AWS

Deploy to Amazon Bedrock AgentCore (Recommended)

Deploy your MCP server to production in minutes:

# 1. Create ECR repository
./create-ecr-repo.sh

# 2. Update configuration
sed -i 's/ecr: auto/ecr: aws-pillar-mcp-server/' .bedrock_agentcore.yaml

# 3. Deploy to AWS
agentcore launch

# 4. Test your deployment
agentcore invoke '{"prompt": "find unused resources in us-east-1"}'

📖 Complete Deployment Guides:

Local Development

# Install dependencies
pip install -e .

# Run locally
python -m aws_finops_mcp

🎯 Quick Overview

  • 76 Tools across 14 categories for comprehensive AWS optimization
  • Category Filtering - Load only the tools you need (NEW!)
  • Dual Modes - stdio for direct integration, HTTP for remote access
  • Cost Savings - Identify unused resources and optimization opportunities
  • Security & Compliance - Find unencrypted resources and security issues
  • Performance Analysis - Analyze and optimize application performance
  • Ready-to-Use IAM Policies - Get started in minutes

📊 View Architecture Diagrams - Visual system architecture and data flows

🆕 What's New

Category-Based Tool Filtering

Problem: Loading all 76 tools can be slow and overwhelming for MCP clients.

Solution: Use MCP_TOOL_CATEGORIES to enable only the categories you need!

# Load only cost and cleanup tools (25 tools instead of 76)
export MCP_TOOL_CATEGORIES="cleanup,cost"
python -m aws_finops_mcp

# 67% reduction in tool count, faster loading, easier navigation

Benefits:

  • 67-89% faster loading for focused use cases
  • 🎯 Better organization - see only relevant tools
  • 🔧 Flexible - change categories without code changes
  • Backward compatible - defaults to all tools

📖 See TOOL_CATEGORIES.md for complete guide

New Tools Added

  • Network: NAT Gateways, VPC Endpoints, Internet Gateways, CloudFront, Route53
  • Storage: S3 buckets, storage class recommendations
  • Containers: ECS clusters/services, ECR images, launch templates
  • Messaging: SQS queues, SNS topics, EventBridge rules
  • Database: DynamoDB tables and utilization
  • Monitoring: CloudWatch alarms and dashboards
  • Performance: Lambda cold starts, API Gateway, DynamoDB throttling, RDS insights, CloudFront cache
  • Security: Unencrypted resources, public S3 buckets, permissive security groups
  • Governance: Untagged resources, tag compliance, cost allocation
  • Capacity: ElastiCache, ECS services, Lambda utilization
  • Upgrade: Lambda runtimes, EC2 generations, EBS types, RDS/ElastiCache engines, EKS versions
  • Cost: Savings Plans, Reserved Instances, EBS optimization, snapshots, data transfer, NAT Gateway

📊 Tool Categories

Category Tools Description
🧹 Cleanup 9 Find unused resources to delete
💰 Cost 16 Cost optimization and analysis
📊 Capacity 9 Resource utilization and right-sizing
🔒 Security 5 Security compliance checks
Performance 5 Performance analysis and tuning
🔄 Upgrade 8 Outdated resource detection
🌐 Network 5 Network resource optimization
💾 Storage 2 Storage optimization
📦 Containers 4 Container resource management
💬 Messaging 3 Messaging service cleanup
🗄️ Database 2 Database optimization
📈 Monitoring 3 Monitoring resource cleanup
🚀 Application 2 Application health monitoring
🏛️ Governance 3 Tagging and compliance

Total: 76 tools - Use category filtering to load only what you need!

# Load only cost and cleanup tools (25 tools instead of 76)
export MCP_TOOL_CATEGORIES="cost,cleanup"
python -m aws_finops_mcp

📖 See TOOL_CATEGORIES.md for complete documentation

Features

76 Tools Across 14 Categories - Use category filtering to load only what you need!

🧹 Cleanup Tools (9 tools)

Find unused AWS resources to reduce costs:

  • `find_unused_lambda_function

Frequently Asked Questions

What are the key features of AWS FinOps MCP Server?

76 tools across 14 categories for AWS resource management. Category-based tool filtering for 67-89% faster loading. Support for both stdio and HTTP modes. Comprehensive cost savings and security compliance analysis. Optimized for deployment on Amazon Bedrock AgentCore.

What can I use AWS FinOps MCP Server for?

Reducing monthly AWS spend by identifying and terminating unused resources.. Automating security compliance checks to find unencrypted storage and permissive security groups.. Right-sizing infrastructure capacity based on actual utilization metrics.. Monitoring application performance and identifying bottlenecks like Lambda cold starts.. Ensuring governance by identifying untagged resources and non-compliant configurations..

How do I install AWS FinOps MCP Server?

Install AWS FinOps MCP Server by running: pip install -e . && python -m aws_finops_mcp

What MCP clients work with AWS FinOps MCP Server?

AWS FinOps MCP Server works with any MCP-compatible client including Claude Desktop, Claude Code, Cursor, and other editors with MCP support.

Conare · memory for coding agents

Turn this server into reusable context

Keep AWS FinOps MCP Server docs, env vars, and workflow notes in Conare so your agent carries them across sessions.

Set up free$npx conare@latest