Introduction

Atlas is Northwind Cloud's product analytics API: you send it events (checkout_completed, feature_used, whatever actions matter to your product), and it gives you funnels, retention cohorts, and ad-hoc queryable event data back. It's built for teams who want analytics infrastructure without running their own event pipeline.

What Atlas is not

Atlas is an events API, not a full customer data platform — it doesn't do CDP-style reverse-ETL, warehouse syncing, or destination fan-out. If you need those, pair Atlas with Cascade for pipeline/warehouse work; Atlas focuses on being the fastest path from "user did something" to "query it in a funnel."

Key features

• Real-time event streaming with sub-second ingestion (p99 < 400ms from POST to queryable)
• Custom funnels and retention cohorts, computed on read — not batch jobs you wait on
• SQL-based ad-hoc querying over raw event data via the query endpoint

Before you start

Read Concepts first — specifically how Atlas models events and identities — before jumping to Getting started. Ten minutes there will save you from the single most common integration mistake: inconsistent event naming that silently breaks funnels weeks later.