Autonomous testing

Let AI investigate.
Make every verdict earn your trust.

Qarmiq’s autonomous agents test scenarios in a real browser. A separate AI judge evaluates the recorded evidence and reports whether the behavior passed, failed, or remains inconclusive.

See Qarmiq in action All features
The real thing

A trace, and a verdict that has to cite it.

The agent’s steps are recorded, and the judge can only rule from what is on the record.

  • An answer that cites no evidence counts as inconclusive
  • Evidence includes a step-by-step trace and a video with typed text masked
  • Checks marked for human verification are skipped

Where it stands. Chrome-family browsers only today. A person starts each run.

Agent traceChrome · real browser
  1. Opens the checkout with a test card set to decline once
  2. Retries with a valid card. Payment confirmed, order created
  3. Opens the new order and looks for its invoice
  4. No invoice is present
Judge: failedCites steps [2] [3] [4]
failed
With nothing cited, the same answer would be marked inconclusive and handed to a person.

Illustrative example.

In practice

What it does.

01

Give the agent a behavior to verify.

Start with a scenario or a run. The agent uses the existing product context to carry out the check and records the steps it takes.

02

A claim isn’t proof.

The judge evaluates the evidence independently of the agent that performed the test. If a verdict cannot cite sufficient evidence, the result is inconclusive rather than an unsupported pass.

03

New discoveries become opportunities to improve coverage.

Agents can propose new checks and learnings based on what they discover. Your team reviews those proposals before they become part of the ongoing verification process.

Why two models

The agent that did the work never grades it.

Code written by an AI should not be graded by the same AI, so Qarmiq does not grade its own work either.

Evidence, or inconclusive

A pass with nothing to show for it is never a pass. It is marked inconclusive and handed to a person.

It proposes, never decides

Tests propose learnings, test data and new checks, and they wait on the Decision Board for a person to accept.

A pass becomes a script

A passing autonomous test produces a Playwright script, which is proven before it counts.

Where it fits

It works with the rest.

Verified scripts

Turn verified behavior into repeatable checks.

A passing autonomous test produces a Playwright script. It is run several times, plus a deliberately broken variant, before it counts as verified, and it updates itself when the product still works.

See Verified scripts → Decision Board

Let AI discover more. Keep your team in control.

Everything Qarmiq proposes waits here, one question at a time, for someone to accept or dismiss: learnings, new checks, route changes and merges.

See Decision Board → Witness

Make every manual test more useful.

A small widget in your own app. Testers test as they normally do, record pass or fail and file defects in place, while Witness keeps the evidence of everything they do.

See Witness →

See an autonomous test investigate a real scenario.

We set up the runner with you and show you a verdict with its cited evidence.

See Qarmiq in action All features

Qarmiq is not self-serve yet, so we set each team up directly. Pricing is one flat price per project.