
- Monthly visits
- 309
- 2026-09
Smoketest - AI QA testing for every ticket and surface
Smoketest reads tickets, comments, files, and linked pull requests when work reaches QA, then tests the affected surfaces in separate sandboxes. It covers web and mobile apps, APIs, MCP servers, and AI agents, and posts results with recordings and steps back to the ticket. Teams can use its test history for regression runs on deploys.
- AI Coding
- Test Automation
- Software
- AI QA ticket testing
- automated cross-surface testing
- QA regression testing




Capabilities
Ticket-based QA testing
Tests work from tickets and reports results with evidence on the ticket. It reads ticket comments, files, linked pull requests, and the team's playbook, then selects checks and the required test accounts.
Multi-surface testing
Tests the surfaces affected by a ticket in separate sandboxes, concurrently. Supported surfaces include web and mobile apps, APIs, MCP servers, and AI agents; each surface runs checks suited to its interactions.
Test recordings and steps
Attaches evidence to each test result so a team can review what happened. Results can include recordings, steps, reasoning, and details of what was not tested.
Ticket-based regression tests
Uses passed ticket tests as regression tests on each deploy. The team can review results and retains ownership of release sign-off.
Pricing
Smoketest offers paid plans, billed monthly or annually for Pro and with custom terms for Scale. Compare monthly run volume, projects, concurrent runs, history, and extra-run arrangements to choose a plan.
| Plan | Price | Included features and limits |
|---|---|---|
| Pro | $399 a month, or $4,000 a year |
|
| Scale | Custom pricing, designed with you |
|
Plans are paid by card through Polar. The pricing FAQ says to contact the team on a call if you need to pay by wire. Cancel from billing; access continues until the end of the paid period.
Use cases
Review tickets waiting in the QA workflow
When a software change reaches the team's QA status, Smoketest reads its ticket and linked material, tests the affected surfaces, and reports a result and evidence on the ticket.
Check a change across multiple product surfaces
When one ticket affects more than one surface, such as a web app and an API, Smoketest runs the relevant checks in separate sandboxes at once and posts the outcomes together.
Run ticket tests as deployment regressions
After ticket tests pass, teams can use them as regression tests on every deploy. Results include evidence for review, while a named person keeps the release decision.
Limitations
A test run lasts up to 10 minutes; longer tests count as additional runs for each further 10 minutes, up to 6 runs at the 60-minute limit.
Monthly plan runs do not roll over; runs purchased in packs last 12 months.
A new attempt uses runs when a ticket is moved back to QA; a repeat within the same attempt is not charged.
The pricing page lists Jira Cloud and Linear as a good fit and says Jira Data Center is not a fit yet.
Private-network or customer-server setups require a custom setup on a Scale plan.
The pricing page says teams that need SOC 2 or another certificate are not a fit yet.
Who it is for
Software teams without a dedicated QA team that need tickets checked before release
QA teams that need a first-pass check across tickets and product surfaces
Engineering leaders responsible for release checks across multiple surfaces
Product managers who need ticket test outcomes visible in their team's workflow
FAQ
Does Smoketest replace QA engineers?
No. It takes a first pass on tickets so QA engineers can focus on exploratory testing, edge cases, and sign-off.
How does Smoketest learn our QA process?
It reads your statuses, tickets, and team-written playbook. Each result shows which playbook rules it applied, and it remembers product knowledge as it tests.
How does it decide what to test?
It reads the ticket, comments, files, and linked pull request, then identifies the affected surfaces and the sandbox needed for each check.
What happens when a test is unclear?
Smoketest reports that it is unsure and leaves the ticket in QA with a note about what was missing. A human status change takes precedence.
Which surfaces can Smoketest test?
It supports web apps, iOS and Android apps, APIs, MCP servers, and AI agents. A ticket can run across multiple surfaces at once.
What is a run, and how long does it last?
A run is one test lasting up to 10 minutes. Each further 10 minutes uses another run, up to 6 runs at the 60-minute limit; tickets waiting in the queue use no runs.
Comments