QA automation · writes · runs · fixes

It writes, runs and
fixes your tests.

From requirement to result, in the browser and in the systems no one else reaches. No CI, no automation team.

👀 it watches your cursor, and shifts color as you scroll

Writes

from what you already have: a ticket, a document, a single paragraph.

Runs

in a real browser and a real desktop app, with video and evidence.

Fixes

when a selector breaks, it rewrites itself and keeps going.

A test management tool records what happened.
Qymar makes it happen.

The proof

Watch the self-repair happen.

The site changed. A button moved. The test no longer wrecks your day.

interactive · click to see it

Failed → fixed → passed.

When an element changes, it rewrites the locator on its own, runs again and keeps going, and records exactly what changed, for you to review.

Only something that's there when it breaks can fix it.

1Open /checkout and add an item
2Fill in the shipping address
3Click
4Check the total with shipping
1,284
tests fixed themselves this month.
example from the dashboard · the number that proves value month after month

Proof it really ran.

Video and screenshots built in, in a single file that opens on any machine with nothing to install. You forward it to your manager and you're done.

How it starts

You pick the ticket.
It does the rest.

No spec to write, no spreadsheet to fill in, nothing to prep.

live now
1

Point it at the source

a ticket, a document, or just a pasted paragraph.

2

It reads and connects the dots

it pulls what's scattered around: the description, the discussion comments, the acceptance criteria on the parent item, the attachment.

What makes a ticket testable is almost never in the ticket. It's in everything around it.
3

Tests, ready to run

the scenarios, plus scripts that actually run.

No requirement? Start with whatever you've got.

Every rung below works on its own. You're never blocked because a doc is missing.

No written requirement?paste the ticket, the email, the WhatsApp paragraph.live now
No ticket at all?answer 4 questions. It writes the rest.in progress
Nothing but the app itself?record your screen using it. It describes what it saw.in progress
Whatever your starting point is, it works.

And it never makes things up.

It never invents a rule nobody wrote down.
Every rule shows where it came from, the field, the comment, the attachment page.
An incomplete requirement blocks nothing: it generates what it can and shows you what was missing.
coverage · regra → cenárioin progress
BR01validate a valid ID number2 scenarios
BR02block an already-registered ID1 scenario
AC01show the error message in red
EF01handle the lookup service timeout1 scenario

It proves what it covered.

Every rule in the requirement is traced to the scenario that tests it. Whatever's left out shows up, with a button to fix it.

"the AI wrote tests", anyone can say that. "the AI proved what it covered" is what no one else shows.
What already works today

Live now: paste the text or upload the document → ready scenarios and tests, with self-repair, in the browser and on legacy desktop.

In progress (and you, as a pioneer, help set the order): pulling straight from Jira / Azure DevOps, the 4-question interview, screen recording, and the coverage matrix.

What no one else does

That system nobody wants to test? It tests it.

Old desktop apps, the kind where a QA still clicks through by hand, one field at a time, every release. It maps the screen, writes the test and runs it, with video of what it did.

No documentation? It sweeps the screen and builds the map on its own.

WinForms · old .NET · the world stuck in banking, industry and government
Legacy System - Registration
Name
ID number
Status
mapping the screen...
Integrations

It works with no cloud at all.
And connects to yours, if you have one.

No feature requires an integration. Your credentials and your data never leave your machine.

Starting fromTest casesRunsBugs
I don't use any of these toolsavailableavailableavailable
Azure DevOps / TFSavailableavailableavailable
Jirain progressin progressin progress
available in progress

The capabilities are composable: run and store everything locally and send only the bugs to your tracker, for example. Nothing here requires a cloud account.

Why Qymar

What the category doesn't do.

CapabilityQymarTest managementOpen frameworks
Actually runs the test, end to endyesnoyes
Fixes the selector on its own, the moment it breaksyesnono
Test legacy desktop (WinForms/.NET)yesnono
Runs locally / data stays inyescloudyes
Starts from a thin ticket or no specyespartialno
Transparent AI cost (your key / free)yesbundledmanual

When all you need is to organize cases and reports for a large team, a test management tool can be the right call. When you need to run the test, it's here.

Pricing

Automation without the old-way cost.

The price follows your team instead of exploding with it.
The Team plan covers up to 5 people for $59/mo: $12 per user instead of the $19 solo seat.
More than 5? Each extra seat is $9.
You pay the AI straight to the provider, at cost, pennies per test, with no markup from us. And re-running a test that already exists costs no AI at all.
Every new account starts with 7 days free. The first 15 people on the pioneer list start with 30.
You don't need an automation team to have automation.
Early access

Join the pioneers.

Limited seats: the first 15 people on the list get 30 days free at launch. Every new account starts with 7 days. You're among the first to run it when it opens, and you help shape v1.

Solo · 1 user · $19/mo Pioneer · 30 days free · 15 seats Team · up to 5 users · $59/mo
✓ You're on the list. I'll ping you the moment it opens.
qymar.com · launching soon