Today we’re announcing the CloudZero Claude Code Plugin, a new capability that puts CloudZero’s full cost intelligence model directly inside Claude Code, where engineers and technical FinOps practitioners already work.
The plugin connects a Model Context Protocol (MCP) server and nine pre-packaged investigation skills to CloudZero’s cost data, covering cloud and AI spend across AWS, GCP, Azure, Snowflake, MongoDB, OpenAI, Anthropic, and more. Engineers can now ask cost, anomaly, and margin questions in plain language and get immediate, context-rich answers — without leaving their workflow, opening a dashboard, or filing a request with FinOps.
The Problem: Cost Intelligence Has Fallen Behind
Here’s the reality of how most engineering teams handle cost today. Something spikes. Someone notices. A Slack thread starts. A FinOps analyst pulls data from one tool. An engineer pulls context from another. Two days later, there’s a hypothesis. Maybe. Meanwhile, whatever caused the spike is either still happening or already forgotten.
This workflow with separate tools, manual exports, cross-team coordination wasn’t fast even when it was the only option. And now, engineering teams are moving faster than ever, driven by agentic workflows and AI-accelerated dev cycles. Using a cost management platform like CloudZero provides powerful insights into the impact of architectural decisions. But they also want to speed up and automate workflows wherever possible.
Cost intelligence needs to be where decisions are made.
The Shift: Cost Intelligence Moves Left
Engineers are already doing development, infrastructure decisions, and increasingly cost analysis inside AI-native environments like Claude Code. They’re connecting data sources like GitHub, Snowflake, Salesforce and getting cross-system answers in a single conversation. Some are dispatching agents that discover issues, analyze across systems, and implement fixes without ever opening a separate UI.
CloudZero is built for this model.
CloudZero’s new Claude Code plugin shifts cost intelligence left into the daily engineering workflow, where architectural and AI spend decisions are actually made. Inside Claude Code, engineers can ask complex cost questions in plain language, investigate spend anomalies, understand cost by owner or service, and get to the root cause without building a single dashboard or manually pulling data from a separate tool. Multi-step analysis — spike → cause → impact → next action — happens in a single conversation.
The result is faster decisions, fewer costly mistakes, and cloud and AI spend that’s optimized continuously, not reconciled after month-end.
Why CloudZero, Not Just Any MCP Integration
We should acknowledge something directly: we’re not the only company to launch a cloud cost MCP integration. But MCP is just a connection standard — like USB. The differentiator is what’s on the other end of that connection.
LLMs reason well. They struggle with financial normalization, cost allocation, and multi-vendor billing complexity. An AI reasoning over raw billing exports, unallocated shared costs, and mixed vendor invoices isn’t performing analysis; it’s guessing. And guesses about cloud and AI costs at scale are expensive.
CloudZero provides the structured, allocated, dimensionalized cost data that makes AI-driven analysis trustworthy. Everything that differentiates CloudZero’s platform — custom dimensions, allocation logic, unit cost modeling, cost per customer, cost per feature, cost per anything — applies fully in the MCP context. The CloudZero Claude Code Plugin doesn’t just connect Claude to cost data. It connects Claude to your cost data, organized the way your business actually works.
This is the difference between an AI tool that gives you an answer and one that gives you the right answer.
Nine Skills. Zero Dashboards Required.
The plugin ships with nine pre-packaged investigation skills that cover the full range of daily FinOps work. Engineers and FinOps practitioners don’t have to know what to ask or how to model the data – the data, context, and workflows are built right in. A few representative examples:
- Cost Anomaly Detection and Investigation — scans for unusual spending patterns and signals of waste, misconfiguration, or optimization, then drills into sudden cost increases and decreases to surface root cause, accountable owner, and recommended next steps
- Top Cost Drivers — identifies and ranks the highest cost contributors across services, accounts, teams, regions, and custom dimensions to prioritize optimization
- Cost Trend Analysis — analyzes cloud cost trends to surface spending patterns, growth rates, and seasonality, enabling accurate forecasting and budget planning
- Margin and Unit Economics Analysis — easily pairs CloudZero cost intelligence with revenue data to calculate cost per customer, feature, or workload, giving teams a fast answer to whether cloud and AI spend is generating profitable returns
That last one matters more than it might initially seem. When we talk about cost intelligence, we’re really talking about both sides of a coin. Knowing your cloud bill is $500K a month tells you the total, but it doesn’t tell you where it went — which customers, which features, which AI workloads are driving the number. The CloudZero Claude Code Plugin lets teams answer the harder question: who’s expensive and why?
What This Looks Like In Practice
“I don’t exaggerate when I say it’s made me 10 times the FinOps practitioner I was before it,” says Larry Advey, Director of Cloud Platform and FinOps at CloudZero. “I can bring in context from any source — engineering events, pull requests, systems metadata, revenue data, cost intelligence metrics — and understand in minutes what changed, why, and what to do about it. Investigations that used to take days now take minutes.”
For a team facing a cost anomaly, the workflow now looks something like this: ask Claude Code what changed and why, get an answer that spans services, teams, environments, pipelines, ticket systems, and custom dimensions, see who’s accountable, understand whether the spend was justified, and decide on next steps — all in one conversation, without leaving the terminal.
That’s not a marginal improvement. It’s a different way of working.
Built For The Agent Era
Engineers are moving cost analysis into AI-native workspaces. CloudZero doesn’t replace that workflow. We’re powering it. This is what shifting cost intelligence left looks like. With the CloudZero Claude Code Plugin, any team facing a cost anomaly can instantly assess what changed, why, who’s responsible, whether it was worth it, and how to fix it if necessary.
The agentic era isn’t coming. It’s here. Engineering orgs that thrive in it will be the ones that bring the right intelligence into the right tools at the right moment. Cost awareness shouldn’t lag behind every other part of the engineering workflow. It should be embedded in it.
CloudZero’s Claude Code Plugin makes that possible today.
Get Started
The CloudZero Claude Code Plugin is available now to all CloudZero customers. A Claude Code subscription is required. Learn more and get started.
Also, register for our webinar on April 2, 2026, to learn how AI is changing engineering workflows, what cost intel looks like in that new model, and what the future will bring.
The Cloud Cost Playbook
The step-by-step guide to cost maturity
