Introducing Graphene Cloud

Today, we’re announcing Graphene Cloud, the hosted complement to your Graphene project.

When we launched Graphene in May, our thesis was simple: to make data more accessible to everyone via agents, you first have to make agents good at data analysis. We reinvented data tools with agent ergonomics in mind: Everything as code and CLIs, a semantic layer designed for legibility and token efficiency, dashboards that render instantly as you build them — all of it was in service of giving a coding agent the context and tools it needed to do real data work.

That thesis holds. But it leaves a gap. Coding agents are for the people comfortable with git, the command line, and environment setup. That’s most of the people we’ve talked to who are excited about Graphene — but it isn’t most of the people at their companies who need answers from data.

So we’ve spent the months since launch on the second half of the problem: making analytics agents, and the reports they produce, accessible to everyone — no matter how technical. That’s Graphene Cloud.

Agentic analytics for all

Graphene Cloud is designed to be the natural complement to a Graphene project you already have in your repo:

• Merge to main, and it’s live. Every page in your Graphene project is automatically synced and hosted on Graphene Cloud as soon as it lands on your default branch.
• Ask and publish from your browser. Not everyone works in your repo, and they shouldn’t have to. The Graphene web agent lets anyone create and publish new pages directly from the browser — describe what you want, and the new page appears. Publishing is as easy as clicking a button, and Graphene Cloud handles the git operations for you.
• Ask questions where you already work. The Graphene Slack agent brings the same semantic layer into your channels, so people can get a real, grounded answer without leaving the conversation they’re already having.

The result is that everyone in your company can ask, create, and collaborate on data — and because it’s all built on Graphene, every answer is backed by the same trustworthy, version-controlled semantic layer your engineers use, not a one-off query improvised from scratch.

In the words of one of our design partners, Atlas:

“Graphene meets our data science and engineering teams inside the agents we already work in — Claude Code and Cursor, local and cloud — and it has made building and sharing analysis dramatically faster. Graphene Cloud turns that into something the whole company owns: hosted pages anyone can open, comment on, and build from, delivered wherever people already are, whether that’s the browser, Slack, or their own agent.”

— Yousuf, Data Scientist at Atlas

Available today

With Graphene and Graphene Cloud together, you get frontier agentic analytics, version-controlled in your repo, without asking your non-technical colleagues to become engineers to benefit from it.

We’re opening up Graphene Cloud for anyone, starting today. Drop us a line to check it out — we offer free trials and free implementations (yes, really).