I use a coding agent

QualityMax gives a coding agent an independent verification loop: discover a journey, turn it into an executable test, run it, and inspect the resulting evidence. Start with the route that matches where you work.

Connect through MCP

The MCP quickstart walks through connecting a supported coding agent to QualityMax. Its workflow is intentionally small: crawl a safe application journey, generate a Playwright test, run it, then inspect the artifacts. Use it when your normal work happens in an MCP-capable editor or agent.

Treat the generated test as a reviewable artifact. Read the intent, selectors, assertions, and result before treating it as coverage for a behavior that matters. The core concepts explain how test cases, automation scripts, executions, and evidence fit together.

Use qmax-code from a terminal

qmax-code is the interactive terminal path. It is useful when you want to start from a local checkout, examine a test run, or keep the feedback loop beside your source changes. The CLI quickstart distinguishes this interactive experience from the local agent that executes tests on infrastructure you control.

Whether you use MCP or qmax-code, keep the same boundary: QualityMax verifies a proposed behavior; your repository and review process remain the source of truth for the change.

Understand personal coding seats

The product map shows where personal coding seats fit in QualityMax. Personal coding seats explains when an individual can use an existing Claude Code or Codex subscription, what that does and does not change about billing, and why one user’s seat turns are serialized. It also distinguishes the currently available connectors from planned options, so confirm availability with your team before planning a workflow around a seat.

Suggested first session

1. Follow the MCP quickstart or CLI quickstart.
2. Choose one stable user journey and generate one focused test.
3. Run it and inspect its evidence before using it in a pull-request gate.