Copilot or autonomous

Autonomous testing, with a person where it matters.

Both modes run the same pipeline: requirement, generate, run, report. The only difference is where a person signs off, and you pick per run.

Checkout must accept coupons
Login flow
Checkout flow
Payment API

3/3 passed

Copilot or autonomous

Decide how much you hand over.

Both modes run the same pipeline. The only difference is where a person signs off, and you can pick per run.

Copilot mode

Review every test case before anything runs.

  1. Requirement
  2. Generate
  3. You review
  4. Run
  5. Report

Best forNew features, regulated flows, teams building trust in AI-written tests.

Autonomous mode

Generate, run and verify end to end. You review the results.

  1. Requirement
  2. Generate
  3. Run
  4. Report
  5. You review

Best forRegression on stable flows, runs triggered by deploys, overnight suites.

In practice

Start in copilot, graduate to autonomous.

New features in copilot

Review the generated test cases while the requirements are fresh. Reviews often expose gaps in the spec itself.

Stable flows in autonomous

Regression on flows that should keep working, runs triggered by deploys and overnight suites.

Same report either way

Pass rate, per-test steps and screenshots, and a reason for every failure with the requirement it affects.

FAQ

Common questions.

In copilot mode your team reviews the generated test cases and chooses which ones run. In autonomous mode testdart generates, runs and verifies the tests on its own, and you review the results afterwards.

1 requirement → test cases → a run → a report

Your next release, tested before it ships.

Point testdart at one requirement you care about. You'll have test cases, a real-browser run and a report in a single sitting.