Home

Build With Gaia

The complete developer guide for building enterprise AI applications on Cohesity Gaia. Go from idea to production — with real code, a Python SDK, and AI-assisted development.

Get Started SDK Reference Examples

8

Chapters

30+

Code Examples

6

Starter Projects

1

Python SDK

The Guide¶

Introduction¶

Understand what Gaia is, how its RAG pipeline works, and the architecture of a Gaia-powered application.

Start here →

Getting Started¶

Authenticate with the API, make your first RAG query, and explore datasets — in under 10 minutes.

Get started →

Core Concepts¶

Deep dives into datasets, RAG queries, conversations, exhaustive search, and document upload.

Learn concepts →

Building Your App¶

Scaffold a full-stack application with FastAPI, React, session auth, and streaming responses.

Build an app →

Advanced Topics¶

Streaming SSE, refine & feedback, dataset discovery, sensitive data, and helper LLM pipelines.

Go deeper →

Deployment¶

Environment configuration, Docker containers, and a production-readiness checklist.

Deploy →

AI-Assisted Dev¶

Set up Cursor IDE with Gaia-specific rules and use an AI-first development workflow.

Use AI →

Marketplace Deployment¶

Package and deploy Gaia apps to the Cohesity App Marketplace — no external hosting required.

Deploy to Marketplace →

5-Minute Quickstart¶

Go from zero to a working Gaia query in three steps.

1. Install the SDK

Bash

pip install httpx pydantic python-dotenv

2. Set your credentials

Bash

export GAIA_API_KEY="your-api-key-here"
export GAIA_BASE_URL="https://helios.cohesity.com/v2/mcm/gaia"

3. Ask Gaia a question

Python

import asyncio, os, httpx

async def main():
    async with httpx.AsyncClient(
        base_url=os.environ["GAIA_BASE_URL"],
        headers={"apiKey": os.environ["GAIA_API_KEY"]},
    ) as client:
        # List datasets
        datasets = (await client.get("/datasets")).json()
        first = datasets[0]["name"]

        # Ask a question
        answer = (await client.post("/ask", json={
            "datasetNames": [first],
            "queryString": "Summarize the most recent documents.",
        })).json()

        print(answer["responseString"])

asyncio.run(main())

Expected output

Text Only

The most recent documents cover the Q4 infrastructure migration,
including server provisioning timelines and the updated disaster
recovery procedures...

Ready for more? Install the Python SDK for a richer developer experience, or jump straight to Your First API Call.

Example Projects¶

Project	What it does	Complexity
Hello Gaia	Minimal CLI that lists datasets and asks a question	Beginner
Chat App	Full-stack chat with streaming responses	Intermediate
Document Search	Exhaustive search, refine, feedback, discovery	Advanced
Gaia MCP Server	Streamable HTTP MCP tools for Cursor/Claude clients	Advanced
Marketplace Chat	Single-container chat app for Cohesity Marketplace	Intermediate
Marketplace MCP	MCP server on the Marketplace — expose Gaia to any AI tool	Advanced

Tech Stack¶

Layer	Technology	Why
Backend	Python + FastAPI	Async, fast, great Pydantic integration
Frontend	React 18 + Vite + TypeScript	Modern SPA with type safety
SDK	`gaia_sdk` (this project)	Typed async client with streaming support
Deployment	Docker + Azure Container Apps + Cohesity Marketplace	Containerized, scalable, deploy anywhere
AI Tooling	Cursor IDE + custom `.mdc` rules	AI generates Gaia-compatible code out of the box