Not long ago, adding a new cost connector to CloudZero was a serious undertaking. We’d task multiple engineers, build in extended review cycles, run a private preview period.
But a single connector could take up to two months from kickoff to customer hands. For the major cloud providers, that timeline was acceptable. The size of the investment matched the scale of the integration.
But the tools landscape has changed. Our customers’ teams don’t just run on AWS and Azure. They’re also running on ClickHouse, Fastly, Twilio, Confluent, Snowflake, and dozens of others, each generating real, variable infrastructure costs that FinOps teams are expected to track and attribute.
When a connector takes eight weeks to ship, that’s not optimal. So we decided to change the economics of that problem, and deliver value to our customers faster, using AI.
One person. Three to four days.
Using Claude-assisted development, we’ve rebuilt how connectors get built. The new framework targets three to four days per adapter, with one engineer, not a team. Our most recent adapter, ClickHouse, took a single day. That was our 16th formal connector. And we plan to keep shipping quickly.
This is a major change. We’ve altered the cost structure of a decision. When a connector costs a team weeks of effort, you need a heavy prioritization framework to justify building it.
When it costs one person three to four days, the framework gets a lot simpler: Is someone asking for it? Can we build it? Build it.
We now have 16 formal connectors covering the cost sources where the bulk of cloud and AI spend lives: AWS, Azure, GCP, Anthropic, OpenAI and more. We’ve just added (or in a couple of cases, improved) seven: Oracle Cloud Infrastructure, Snowflake (now with tag support, Confluent Cloud, Fastly, Github Enterprise, Twilio, and ClickHouse Cloud.
Connector breadth has become a big factor in how customers evaluate cost intelligence platforms. We’re thrilled to be able to ship connectors quickly, expanding what’s possible and adding another reason that makes CloudZero the most attractive provider to prospects dealing with a growing number of cloud providers.
Keep in mind, too, for those cloud services for which we don’t yet have a formal connector, CloudZero’s AnyCost framework remains a powerful differentiator. AnyCost enables your teams to build adaptors for additional cost sources.
It extends our intelligence layer to any cost source that generates cost data: other cloud platforms, custom AI providers, internal chargeback systems, niche SaaS tools, or emerging inference platforms that don’t yet have native connectors.
The result is a single system of record for cost-to-serve, regardless of where your spend occurs.
Research Report
FinOps In The AI Era: A Critical Recalibration
What 475 executives told us about AI and cloud efficiency.
What’s shipping – and how we do it fast
With our latest connector, ClickHouse, CloudZero customers can now get full visibility into compute, storage, backup, data transfer, and ClickPipes costs, with tag-based allocation through CloudZero CostFormation.
We have at least 10 more additional connectors in the development queue, tied to active customer requests and growing presence in the market.
The reason we can build connectors this fast isn’t just AI-assisted development, it’s what we’re building into. Every cost source CloudZero ingests maps to the Common Billing Format (CBF). This normalized schema captures cost, time, resource identity, and business context.
CBF is simple; it’s a flat, well-defined target that any billing API can map to. A ClickHouse billing record and an AWS CUR line item look nothing alike at the source, but once they’re in CBF, they’re the same shape, able to be allocated through the same CostFormation rules, visible in the same dashboards, and attributed to the same teams and customers.
That helps make a three-day connector possible. The engineering problem isn’t really building an integration from scratch. It’s “map this provider’s API to a format we already know how to allocate.” A connector spec documents the API surface, the auth model, and the field-by-field CBF mapping.
That spec becomes the input for Claude-assisted development; Claude generates the adapter code, the transformation layer, and the tests. An engineer validates against real data and handles the edge cases that only surface in production: undocumented API behaviors, timestamp inconsistencies, and the CostFormation dimension design that determines what allocation paths the connector unlocks.
CBF also means the dimensional engine doesn’t care where the data came from. ClickHouse costs flow through the same allocation rules as Snowflake or Anthropic costs. A FinOps team writes one set of CostFormation rules and every connector benefits. Each new connector doesn’t just add a row to the integrations page, it adds another cost source to an allocation model that already works.
The benefit for FinOps teams and beyond
We don’t believe that a FinOps team, or engineers, or finance teams, should be stuck with a blind spot every time their organization adopts a new tool. Our goal is to be a platform that moves as fast as the teams we support, one where “we’ll add that connector” means days, not quarters.
The backlog is open. If your team is running something you’d like to see covered, reach out to your FAM (for existing customers) or contact us for a discussion and demo. You can also read more about our connectors here.
FinOps In The AI Era: A Critical Recalibration
What 475 executives told us about AI and cloud efficiency.
