Getting Started with FinOps MCP Resources
May 9, 2026 · View on GitHub
Welcome! This repository is designed to help FinOps practitioners experiment with Model Context Protocol (MCP) servers and turn LLMs into practical FinOps copilots.
If you’re new here, here’s the best path to get started:
1. Start with the AWS Pricing MCP Tutorial
Your first stop should be the AWS Pricing MCP Quickstart. It’s a step-by-step guide that walks you through:
- Installing the AWS Pricing MCP server
- Connecting it to a client
- Running your first real-time cloud pricing queries
This is the easiest way to see MCP in action for FinOps.
2. Set Up an MCP Client
Once you've tried AWS Pricing, explore how to use different MCP clients.
As of May 2026, we provide comprehensive guides for 9 major MCP clients:
For Developers & Technical Teams
- Claude Code - Remote MCP, task workflows, enterprise controls
- Kiro - Agentic IDE with spec-driven development (AWS-focused, preview)
- VS Code + Cline - Flexible, extensible, familiar to engineers
- Cursor - Modern IDE with MCP support
For Business & Finance Stakeholders
- ChatGPT - Most accessible, huge user base (added March 2025)
- Claude Desktop - Fast prototyping, conversational
- Microsoft Copilot - Microsoft 365 integration
- Google Gemini - GCP integration, Vertex AI (added April 2025)
For Cloud-Specific Use Cases
- Kiro - AWS-focused agentic IDE
Not sure which to choose? Check our Client Comparison Guide with detailed pros/cons for FinOps professionals.
3. Review Governance & Security
Before scaling further, check out the security and permissions guidance:
This will help you set up safe, scoped credentials and avoid giving MCP servers unnecessary permissions.
4. Explore FinOps Use Cases
Once your setup is running, try applying MCP to real FinOps challenges:
- Compare EC2 and VM costs across regions
- Generate automated monthly budget variance reports
- Run “what-if” simulations for commitment coverage
- Spot anomalies in billing data
👉 And most importantly: come up with your own use cases!
This repo is a sandbox — test, prototype, and see how MCP can accelerate your FinOps workflows.
5. Discover More MCP Servers
The MCP ecosystem is now enterprise infrastructure: at MCP Dev Summit North America 2026, the Agentic AI Foundation reported 110+ million monthly SDK downloads and 170 member organizations. As of May 2026, here's how to discover FinOps-relevant servers:
Official MCP Registry
- URL: MCP Registry (check modelcontextprotocol.io)
- What it is: Official server directory launched by the MCP community
- Coverage: servers for AWS, Azure, GCP, and third-party FinOps tools
Community Directories
- MCPdb: Claims 10,000+ servers and clients - mcpdb.org
- Awesome MCP Servers: Curated GitHub list - github.com/punkpeye/awesome-mcp-servers
- Official MCP Servers: Anthropic's reference implementations - github.com/modelcontextprotocol/servers
What to Look For
When discovering MCP servers for FinOps:
- Cloud provider integrations: AWS Cost Explorer, Azure Cost Management, GCP BigQuery billing
- Third-party platforms: Vantage, CloudHealth, Cloudability, Apptio
- Governance tools: Tagging compliance, policy enforcement, security scanning
- Data sources: Billing exports, usage metrics, reservation/savings plan data
Next Steps
- Try additional MCP servers beyond AWS Pricing (Azure, GCP, Vantage, etc.)
- Browse the MCP Registry to discover new FinOps-relevant servers
- Share feedback or new use cases in Issues/PRs
- Join the wider conversation at modelcontextprotocol.io
💡 Tip: Start small, validate results against raw billing data, and always apply governance guardrails before scaling MCP in enterprise environments.