Overview

Qassandra is an autonomous QA platform. AI agents drive a real browser to discover the user flows in your app, then re-run those flows on every change to catch real bugs and visual regressions.

Traditional end-to-end test suites (Playwright, Cypress, Selenium) require you to write, maintain, and constantly debug brittle scripts. Qassandra replaces most of that work with AI agents that look at your app the way users do — they take screenshots, understand the UI, and decide what to click next.

How it works

The product has four moving parts that you'll see across the docs:

1. Apps — a web application you want to test. You give Qassandra a URL and it figures out what the app does.
2. Flows — the individual user journeys (e.g. “Log in with valid credentials”, “Add item to cart”) Qassandra discovers and turns into repeatable tests.
3. Checks — a single execution that runs a selected subset of your flows against the live app, captures screenshots, and reports results.
4. Suggestions — newly discovered flows that you can approve or reject so your test suite grows alongside your product.

Why teams use Qassandra

• No tests to write. Discovery bootstraps your first test suite the moment you connect an app.
• Self-healing execution. When a CSS selector breaks, an AI fallback locates the element and updates the step automatically.
• Visual regression built in. Every step captures a screenshot and is compared against the latest baseline on your main branch.
• PR-aware. When triggered from a pull request, Qassandra reads the diff and prioritises flows that are most likely affected.
• Fast first, AI on failure. Steps run with raw Playwright selectors at full speed; AI only steps in when something looks wrong.

Where to go next

If you're here to ship, jump straight to the Quickstart. If you want to understand the model first, start with Apps and follow the Core concepts section in order.