AI Test Generation Agent — Turn User Stories into Test Suites
Automated test case generation from text, images, API specs, and live URLs — for Web, Android, and iOS.
Qgen — Pcloudy's automated test case generation agent and part of the QPilot.AI platform — reads your requirements, mockups, API docs, or live URLs and generates structured, executable test cases for Web, Android, and iOS automatically. No code required.
- TC-01 · Valid email + password → successful login
- TC-02 · Invalid email format → inline error
- TC-03 · Wrong password → error + retry counter
- TC-04 · 5 failed attempts → account locked
- TC-05 · Session timeout after idle period
- TC-06 · Empty fields → mandatory field validation
Generate from whatever you have
Four input types cover every stage of the SDLC — from a one-line user story to a fully built app.
Description (Text)
Paste a user story, feature spec, or functional description. Qgen interprets roles, flows, validations and edge cases.
Upload Image
Drop in a UI screenshot, wireframe or mockup. Qgen reads fields, buttons, labels and flow to generate UI test scenarios.
API Document (JSON)
Upload an OpenAPI/JSON spec. Generate API test cases for request/response validation, mandatory fields and error codes.
Website URL
Provide a live URL. Qgen crawls page structure and UI elements to generate navigation, validation and functional flow tests.
What a generated test case looks like
Not just titles. Qgen produces fully structured cases — preconditions, steps, expected results and test data — ready to execute or export.
- User account exists and is active
- Lockout policy = 5 attempts / 15 minutes
password: WrongPass!23
- Navigate to /login
- Enter valid email + invalid password
- Tap Sign In
- Repeat steps 2–3 four more times
- Attempt a 6th login with valid credentials
- Inline error after each failed attempt
- Account locked after 5th failure
- Valid login on 6th attempt is rejected with "Account locked" message
One agent. Three platforms.
Pick one or more platforms per scenario set. Qgen tailors every generated test case to platform-specific patterns — instantly.
Web
Browser flows, form validations, responsive checks across Chrome, Edge, Safari & Firefox.
Android
Touch gestures, system dialogs, permissions, intents — across OEMs and OS versions.
iOS
Native iOS patterns, biometrics, push, deep links — every iPhone & iPad generation.
One scenario set → generated, tailored, and ready to run on all three.
From input to executable test cases
Five steps. No code. No templates to maintain.
Create a scenario set
Name your scenario set (e.g. Login Functionality), pick platforms, and start.
Provide an input
Paste a description, upload an image or API JSON, or drop a URL. Click Analyze.
Review scenarios
Qgen lists generated scenarios. Select all, deselect noise, or refine with AI feedback.
Generate test cases
Click Analyze again — full structured test cases are created from the chosen scenarios.
Refine & export
Edit inline, refine with AI, then export to your test management tool of choice.
Built for real test coverage
Positive, negative & edge cases
Happy paths, validation failures, boundaries and error states — generated together in one run.
Refine with AI feedback
Add plain-English feedback ("add 2FA scenarios", "cover offline mode") and the entire suite updates.
Visual element analysis
Image input understands buttons, fields, labels and flow — not just OCR — for accurate UI scenarios.
Generate → run on real devices
Built into Pcloudy: take generated cases straight to execution on real Android & iOS devices.
Exports straight into your test management stack
Push generated cases to the tools your QA team already uses — or export structured CSV / Excel / JSON.
Jira / Xray
Zephyr
TestRail
qTest
Azure DevOps
CSV / Excel
JSONAlready have a test suite?
Qgen plays well with the work you've already done — bring your existing cases in and let the agent extend coverage where it actually matters.
Import existing cases
Upload CSV/Excel from TestRail, Zephyr, qTest or Jira/Xray. Qgen indexes them as your baseline.
De-duplicate & merge
Generated cases are matched against your baseline so you don't end up with duplicate scenarios.
Coverage gap analysis
Qgen highlights missing negative paths, boundary cases and untested flows from your requirements.
Why not just paste it into ChatGPT?
General chatbots can draft cases. They can't run them, structure them, refine the suite or hand off to your QA stack.
Your data stays yours
Inputs you provide to Qgen — user stories, mockups, API specs, URLs — are used only to generate your test cases. They are never used to train foundation models. Pcloudy runs on enterprise-grade infrastructure with encryption in transit and at rest, role-based access, and tenant isolation.
Manual test design doesn't scale
- QA spends days authoring test cases by hand for every release
- Edge cases and negative paths get missed under deadline pressure
- Stories, mockups and APIs all need separate manual translation
- Coverage drifts as features evolve faster than the test plan
- Full scenario sets in seconds from text, image, API spec or URL
- Positive, negative and edge cases generated together
- One workflow for stories, mockups and APIs across Web/Android/iOS
- Refine with AI feedback to keep coverage in lockstep with features
Real scenarios
Mobile banking app
API contract testing
Existing web app regression
Questions, answered
What is an AI test generation agent?
An AI test generation agent is an autonomous system that reads product artifacts — user stories, UI mockups, API specifications, or live URLs — and produces structured, executable test cases without manual authoring. Pcloudy's Qgen, part of the QPilot.AI platform, generates positive, negative, and edge cases for Web, Android, and iOS in a single run, and hands them off to real-device execution or your existing test management stack.
How does automated test case generation work?
You provide an input (text, image, JSON, or URL). Qgen parses it with vision and language models, identifies actors, flows, fields, validations, and error states, then drafts scenarios. You select the scenarios you want, and Qgen expands each into a full test case — preconditions, steps, test data, and expected results — ready to export or execute.
How is Qgen different from using ChatGPT for test generation?
ChatGPT produces free-form text. Qgen produces structured, exportable test cases tailored to Web, Android, and iOS patterns; refines an entire suite from a single feedback prompt; pushes natively to Jira/Xray, TestRail, Zephyr, qTest, and Azure DevOps; and runs the generated tests on real devices inside Pcloudy. Your inputs are never used to train foundation models.
Is AI test generation suitable for banking and fintech app testing?
Yes. Qgen is used by BFSI teams to generate test cases for funds transfer, KYC, OTP, account lockout, fraud-flag, and compliance flows. Pcloudy's infrastructure is PCI-DSS, SOC 2 Type II, and ISO 27001 compliant, with encryption in transit and at rest, role-based access, tenant isolation, and no model training on customer data.
How does the AI test generation agent handle negative and edge cases?
Qgen generates positive, negative, and edge cases together in one run — invalid inputs, boundary values, validation failures, timeouts, retry counters, lockouts, and error states. You can also use the AI refine prompt to add specific edge categories (e.g. "add 2FA failures", "cover offline mode").
Can generated test cases run on real Android and iOS devices?
Yes. Qgen is built into Pcloudy, so generated test cases can be taken straight to execution on 5,000+ real Android and iOS devices on real networks — no emulators, no separate setup.
Can it generate Selenium or Playwright scripts?
Qgen produces structured test cases (steps, data, expected results) that can be executed in Pcloudy or used as the source for Selenium, Playwright, and Appium automation. Direct script generation for these frameworks is on the near-term roadmap.
Does it support BDD / Gherkin format?
Yes. You can ask Qgen to format scenarios in Given / When / Then and export them as .feature files for your BDD pipeline.
Which LLM does Qgen use?
Qgen runs on a curated mix of leading foundation models, selected per task (text reasoning, vision parsing for mockups, structured API analysis). Your inputs and generated cases are never used to train those models.
Does it work for API testing?
Yes. Upload an OpenAPI / JSON spec and Qgen generates request/response validation, mandatory field checks, status code coverage, and error-handling test cases — without manual mapping.
