Model Context Protocol (MCP)
The Model Context Protocol (MCP) is an open standard for connecting AI applications to external tools and data. An MCP server exposes tools, resources and prompts over JSON-RPC, and any MCP client, such as Claude, ChatGPT or Cursor, can use them, so one integration works across many AI assistants.
In plain English
Think of USB for AI assistants. Before it, every device needed its own cable for every computer; with one standard port, any device plugs into any machine. MCP does that for connecting assistants to tools and data.
Worked example
| Question | Why did Checkout's cost rise in March? |
|---|---|
| Tool the assistant calls | Cost by service for Checkout, March against February |
| Access used | The engineer's own, read-only |
| Data returned | $22,000, up $2,500 (13%) |
The assistant needed no custom CloudYali plugin: it found the tool through MCP, called it, and answered from live data. The same server works in Claude, ChatGPT, Cursor and other MCP clients.
Why MCP matters
Before MCP, every AI assistant needed its own integration with every tool: N assistants × M tools. MCP turns that into N + M. A tool vendor builds one server, and every MCP client can use it.
Anthropic introduced MCP in November 2024 as an open standard. In December 2025 it was donated to the Agentic AI Foundation, a directed fund under the Linux Foundation, and its specification is developed in the open. Cloud providers publish servers too: AWS's open-source Billing and Cost Management MCP server, for example, gives assistants access to Cost Explorer, Budgets, Cost Optimization Hub and Cost Anomaly Detection.
How MCP works
An MCP host, such as an AI assistant, runs a client for each server it connects to. Servers offer tools, which the model can call; resources, which it can read; and prompts, which users can pick. Messages use JSON-RPC 2.0. Local servers usually talk over standard input and output, and remote servers over Streamable HTTP, with OAuth for sign-in.
Common mistakes
- Giving a server more access than it needs. An assistant can call any tool a server exposes. Prefer read-only servers for data such as costs, and sign in with your own scoped permissions.
- Trusting every server. Tool descriptions and results from a server you don't trust can carry instructions for the model. Use servers from trusted providers, and review what each tool can do.
- Connecting everything. Tool definitions and results are sent to the model as tokens. Dozens of connected servers add input tokens, and cost, to every request.
How CloudYali helps
CloudYali runs a read-only remote MCP server at mcp.cloudyali.io. Add it to Claude, ChatGPT, Cursor or another MCP client, sign in, and ask about costs, budgets, savings, anomalies, tags and resources across every connected cloud and AI tool, with your own permissions.
Related terms
AI, data & SaaS costs
AI cost tracking
AI cost tracking is collecting what an organization spends on AI, from model APIs such as Anthropic and OpenAI, cloud AI services such as Amazon Bedrock, and AI tools such as coding assistants, in one place, and attributing it to the teams, products and models behind it. It applies FinOps to AI spend.
AI, data & SaaS costs
Cost per token
Cost per token is what a language model charges for each token it reads or writes, usually quoted per million tokens, with input, cached input and output tokens priced separately. Output tokens typically cost several times more than input, so the mix of token types, not just the volume, decides the bill.
FinOps foundations
FinOps
FinOps is an operational framework and cultural practice that maximizes the business value of technology, enables timely data-driven decisions, and creates financial accountability through collaboration between engineering, finance and business teams. In practice, the people who spend on the cloud see what it costs, own it, and decide with that cost in mind.
Governance & planning
Cost anomaly
A cost anomaly is spend well outside what a service, account or team normally costs, such as a day costing several times its recent average. Anomaly detection compares each day with a baseline learned from history and alerts on the difference, so a runaway job is caught in days, not at the end of the month.
Governance & planning
Budget
A budget is a planned spending limit for a period, such as a team's monthly cloud cost, with alerts when actual or forecasted spend crosses thresholds you set, such as 80% and 100%. On AWS, Azure and Google Cloud a budget warns rather than stops spending, unless you attach actions or automation to it.
Frequently asked questions
What is an MCP server?
A program that exposes tools, data or prompts to AI assistants through the Model Context Protocol. It can run on your own machine or be hosted remotely, like CloudYali's.
Who maintains MCP?
Anthropic created it and, in December 2025, donated it to the Agentic AI Foundation under the Linux Foundation. The specification is open source and developed in public.
Is it safe to connect an MCP server to my AI assistant?
It's as safe as the server and the access you give it. Prefer read-only servers from trusted providers, sign in with your own scoped permissions, and review which tools an assistant may call.
