Skip to content

Introducing Build With Gaia

There is a specific kind of frustration that every developer building on enterprise AI platforms eventually runs into. You have the API key. You understand the concept. You know roughly what you want to build. But the gap between "I want to ask this AI a question about my company's data" and "I have a working, secure, production-ready application" is enormous — and almost none of that gap is documented anywhere.

That frustration is exactly why Build With Gaia exists.

Why This Site Exists

Cohesity Gaia is genuinely powerful. The ability to index your organization's private data — across object storage, NAS, backup snapshots — and query it with natural language is transformative. It means your AI applications don't hallucinate generic answers from public training data. They reason about your runbooks, your incident history, your financial records, your compliance documents.

But powerful APIs without accessible developer guides produce one outcome: underutilization. The technology sits in the data center, doing less than it could, while developers struggle to bridge the gap between the REST documentation and the running application.

Build With Gaia is our answer to that gap.

What We Set Out to Build

When we started this project, we made a deliberate decision: we would not build a marketing site for Gaia. We would build a developer guide — the kind of resource you actually need when you're sitting at your keyboard trying to make something work.

That means:

Real, runnable code. Every example in this guide starts, runs, and does something useful. We don't show pseudocode or elide the "boring" parts. The boring parts — authentication, error handling, streaming, session management — are often the hardest parts, and we show them all.

A Python SDK. We built gaia_sdk specifically for this project. It wraps the raw REST API with typed async methods, a clean exception hierarchy, and streaming support. When you start a new Gaia project, you don't have to figure out the right way to call the API — you install the SDK and use GaiaClient.

Progressive examples. We start with a single Python script that fits in 80 lines. We end with a full Marketplace-deployable application with a React frontend, SSE streaming, and an MCP server that connects to Cursor and Claude Desktop. You pick your starting point.

AI-first development tools. The .cursor/rules/ directory encodes Gaia-specific development patterns so that when you use Cursor or Claude Code, the AI generates correct code for your project, not generic boilerplate.

The Mission

Our mission is straightforward: make it easy for enterprise developers to bring AI-ready data to their AI applications — safely, securely, and responsibly.

That last part matters. Enterprise AI applications aren't just about capability. They're about:

  • Safety — Your application shouldn't surface sensitive data to the wrong user. Gaia's permission-aware query system, and the patterns in this guide for using it, let you build applications that respect your organization's access controls.

  • Security — Your API keys should never reach a browser. Your users' queries should be proxied through a backend you control. The Backend-for-Frontend pattern at the core of every example in this guide keeps secrets on the server side.

  • Responsibility — AI applications that answer questions from private data need to be transparent about what they know and don't know. Gaia's citation system surfaces sources. This guide shows you how to present those sources to users so they can verify claims instead of just trusting them.

Who This Is For

We built this for the developer who just got access to a Cohesity cluster with Gaia enabled and is trying to figure out what to build with it. We built it for the enterprise architect who needs to explain to their team how to integrate Gaia into an existing application. We built it for the curious engineer who wants to understand how RAG actually works in production, not in a Jupyter notebook.

We also built it for ourselves — because Compound Engineering means every piece of work should make the next piece easier, and a thorough developer guide is one of the highest-leverage investments a platform team can make.

What Comes Next

This guide is a living document. Cohesity Gaia is evolving rapidly — OAuth2 support, expanded indexing capabilities, tighter integration with the broader Cohesity platform. We'll document those changes here as they land.

We also want to grow the community. The Community Projects section is a start — a place to showcase what developers are building on top of Gaia. If you've built something, we want to hear about it.

Build something remarkable.

— Greg Statton, VP & CTO APJ, Cohesity