Introduction

Blunox is an AI Data Automation platform. Autonomous agents plan, build, validate, and deploy production data pipelines — with a human approving the plan at each checkpoint. This guide covers how to go from a connected source to a shipped data product.

Quickstart

Get from zero to a deployed pipeline in four steps:

• Connect a source from the connector library (or generate one from an API spec).
• Open Mission Control and describe the data product you need in plain language.
• Review the generated build graph — sources, transforms, targets, tests.
• Approve. Agents build, test, document, and deploy inside your controls.

Core concepts

Agents

Expert workers mapped to data roles — Builder, Migration, Quality, Activation, FinOps, and Docs. They run concurrently and share the Context Graph.

Context Graph

Your standards, contracts, and policies, encoded once and inherited by every pipeline. See Context Graph.

Mission Control

The console where you supervise agents, review diffs, and approve at checkpoints.

Making requests

Requests can be made in the UI, via the CLI, or through the REST API. Describe the source, target, transformation rules, and business logic; Blunox interprets intent, asks clarifying questions, and confirms understanding before execution.

The build graph

Before any code is written, Blunox produces an explicit graph of sources, transformations, targets, dependencies, and tests. You can edit or reject any node. Nothing builds until the plan is approved.

Approvals

Every change ships through an approval checkpoint as a reviewable diff with tests and lineage attached. Approve to deploy, or request changes in natural language.

CLI & API

Authenticate with blunox auth login, then drive everything programmatically. The REST API mirrors the CLI for CI/CD integration.