Introduction

What TestCatalyst is for

An AI-powered test case management platform built around one idea: QA should not be the thing everyone waits on. It should be the thing that gives everyone confidence, fast, and visibly.

Most test-case tools are a place to store checklists. TestCatalyst does that too, but its actual job is different: it does the repetitive parts of testing, generating cases from a ticket, running them, healing a broken selector, filing and re-closing bugs, drafting a report, so the person on your team who understands the product can spend their time on the parts a model can't do: judging what's actually risky, exploring the paths nobody specified, deciding when "technically passing" still isn't good enough.

The problem this is solving

QA teams get treated as a bottleneck for a structural reason: they're usually the last human checkpoint before release, and everything upstream of them (vague tickets, missing acceptance criteria, flaky selectors, manual re-testing after every fix) eats their time before they even get to the testing that requires judgment. Developers, meanwhile, often only see QA's output as a pass/fail message days later, with no visibility into what was actually exercised.

TestCatalyst attacks both sides of that at once:

  • It removes the grunt work before QA gets involved. Under-specified tickets get bounced back automatically with specific gaps listed, test cases get drafted from the ticket or PRD, and broken selectors repair themselves instead of failing a whole run.
  • It makes QA's work visible while it happens. Every AI-driven run, whether started in TestCatalyst or from a developer's own MCP-connected agent, streams to a live URL. Nobody has to ask "is this tested yet?", they can watch.

The reframe this docs site uses

Instead of splitting into "QA docs" and "dev docs" that never reference each other, most guides here describe the same underlying system from two angles. See For QA engineers and For developers.

How the product is organized

A workspace is built from a small number of concepts that show up everywhere:

ConceptWhat it is
ProjectThe top-level container, usually one per product or codebase. Has suites, a Jira project key, and its own team access list.
SuiteA group of test cases, usually mapped to a feature or ticket.
RunOne execution of a suite: scripted (Playwright), agentic (TestCatalyst's own AI), or via MCP (your own agent).
BugCreated automatically when a test case fails, or manually. Auto-resolves when its test case passes again.
CAT recordingA screen recording, from the bundled Chrome extension, an agent run, or a manual upload.
ReportA branded, shareable summary of a run or a suite, emailed, exported to Confluence, or downloaded.

The full picture, with how these connect, is in Core concepts.

Where to go next