Smoketest
Sign inBook a demo

Ready for QA.
Tested. Done.

Smoketest picks up every Jira or Linear ticket the moment it reaches QA, tests it the way a human would, and moves it to Done. Tickets stop piling up at the end of the sprint.

Book a demoStart free trial
Try it for a real sprint. 14 days free.
WEB › Active issuesBoardListCycle 31 · 3 days leftFilterDisplay
Todo2+
WEB-296
Bulk archive projects from settings
UI2JP
WEB-297
Rate limit public API keys
API3OD
In Progress5+
WEB-284
Apply annual discount at checkout
Billing3SK
WEB-288
Show proration on seat change
Billing2OD
WEB-291
Invite teammate from workspace settings
UI1ML
WEB-279
Recover account with magic link
AuthBug2SK
WEB-293
Export invoices as CSV
Billing3JP
Ready for QA0+
Done1+
WEB-276
Show updated invoice after seat change
Billing2ML
Smoketestpassed
Passed · recording

From Ready for QA to Done,
in five steps.

01
A ticket moves to Ready for QAWe watch the QA column you point us at.
02
We read the ticket, the comments, and the code changeWith GitHub connected, the pull request behind the ticket is read too.
03
We write the test scenariosFrom the acceptance criteria, the way your QA engineer would.
04
We run them like a userDesktop and mobile, with your staging login.
05
Steps, recording, and reasoning land on the ticketPassed moves on. Failed goes back to the developer. Unclear waits for a human touch.
WEB › WEB-288Done
Show proration on seat change
When a workspace changes seat count mid-cycle, the invoice preview should show the proration as its own line.
Activity
Smoketest4 min ago· edited

Passed · 4 of 4 criteria · 11 steps · 1m 40s · desktop

  • Seat count can be increased from 5 to 6
  • A proration line appears in the summary
  • Proration amount matches the remaining days
  • Invoice preview shows proration as its own line

RecordingStep-by-step reasoning

Smoketest moved this issue from Ready for QA to Done

What changes, for each seat

QA engineersMinutes, not daysReady for QA becomes a result in minutes.
DevelopersFix it the same hourStep, recording, assignee. Done.
ProductDone means testedOpen any ticket. See the proof.

Time and money back every sprint

Your team moves 10 tickets to QA each sprint and spends about 2 hours checking each one.

That is 2+ working days, $980, back every sprint.

Manual first pass
20 h
With Smoketest
~2 h review

Valued at $49 per hour, the U.S. median for QA analysts and testers (BLS, May 2024). First pass only; faster bounce-backs come on top.

Everything included

Connect your board. The rest is on us.

Jira and LinearSpecify a QA status. We pick up every ticket that lands there.GitHub PR contextThe code change behind the ticket shapes the scenarios.Desktop and mobileEach scenario runs where your users actually are.Recording on every runEvery run is recorded and linked from the ticket.Logins and environmentsStaging, preview URLs, basic auth, saved sessions.OTP and magic linksInboxes for the flows that need an email to finish.Slack and email alertsFailures and recoveries, per project and channel.Sprint overviewCoverage, first-pass rate, returns, and late QA.

Works with what you already use

JiraLinearGitHubSlackDiscordVercelClaude CodeCursorJiraLinearGitHubSlackDiscordVercelClaude CodeCursor

Shipped recently

All releases →
Aug 23, 2026new
Owner, admin, and viewer, per project
Aug 22, 2026improved
Steadier runs on messy pages
Aug 21, 2026new
Run digests: twice daily to monthly

Questions we get on every call

Does this replace our QA engineers?

No. It makes them faster. Every ticket that reaches QA gets its first pass automatically: scenarios written from the acceptance criteria, run on desktop and mobile, result and recording on the ticket. Your QA engineers stop re-clicking the same flows and writing up the same results, and spend that time on exploratory testing, edge cases, and sign-off. Teams use that room to ship more, not just to spend less.

What do we have to change in Jira or Linear?

Nothing. Pick the status that means ready for QA and we watch it. Your columns, assignees, priorities, and workflow stay exactly as they are. We post one comment per ticket and move it to Done or back to In Progress based on the result. If a person moves a ticket while we are testing, we stop and never move it back.

What if a ticket has no acceptance criteria?

We write scenarios from whatever is there: the description, the comments, the attachments, and, with GitHub connected, the pull request behind the ticket. Most tickets have enough. When one does not, it waits in QA with a note saying exactly what was missing, so the author can add a line and we run it again. Over time that nudges the whole team toward tickets that can actually be tested.

How do you log in to our staging?

You save a login session once per environment and we reuse it. Preview URLs per branch, basic auth, and staging behind a VPN on Enterprise all work. For flows that need an email to finish, each run gets its own inbox, so OTP codes and magic links are handled without anyone touching a phone.

What happens when a test fails?

The ticket goes back to In Progress with the failing step, what we expected, what we saw, and the recording, assigned to the author of the change. The developer fixes it while the branch is still fresh, moves the ticket back to QA, and we test it again. No repro to write, no back-and-forth in the comments.

See all questions and pricing →

The QA column is no longer where the sprint goes to die.

Move one ticket. Watch it come back tested. Then decide.

Book a demoStart free trial