describe('AI agents')
AI agents that test for you: in CI/CD, on your merge requests, in your tickets. They leave the lab one at a time; leave your email on the ones you're interested in to hear when they launch.
When a test breaks because of a selector, it finds the new element and suggests the fix.
The agent analyses a failure caused by an element that can’t be found, locates the element in the new version of the screen and suggests a more robust selector.
Playwright · Maestro · maintenance
It reviews every merge request from a testing angle: missed cases, regression risks, missing tests.
The agent comments on GitLab merge requests with the test scenarios to plan, the risky areas and the missing tests.
GitLab · code review · regression
It reads your OpenAPI contract and generates an API test suite: contracts, error cases, basic security.
From an OpenAPI file, the agent generates contract and error-case tests, then flags gaps between the documentation and the actual behaviour.
API · OpenAPI · contracts
It reads every merge request, picks the tests to run, analyses failures and comments the result in the MR.
Designed for startups without QA, an agent plugged into your GitLab pipeline: it selects the tests affected by the change, reruns failing ones to rule out flakiness, then summarises the likely cause of a failure directly in the merge request.
CI/CD · GitLab · startups
It browses your application like a curious user and reports what breaks, with screenshots and steps.
Given a goal in plain language, the agent explores the application in a real browser, tries unexpected journeys and produces a session report.
exploratory · Playwright · browser
It suggests unit tests for the changed code, runs them and only keeps those that pass and add coverage.
From a function or a diff, the agent generates unit tests, runs them, discards those that fail or add nothing, and opens a proposal ready for review.
unit tests · coverage · TypeScript
It qualifies every new ticket: duplicate, severity, component, missing steps, and suggests an assignee.
Plugged into your ticketing tool, the agent detects duplicates, asks the reporter for missing information and suggests a severity and a team.
Jira · triage · defects
Paste a user story, get structured test cases: happy path, edge cases, errors, in Gherkin or table format.
The agent turns a user story and its acceptance criteria into test cases ready to run or automate, flagging the story’s ambiguities.
user stories · Gherkin · BDD
Got a repetitive testing task that deserves its own agent? Describe it; the best ideas join the lab.
Suggest an idea