Nexus StudioEvery Nexus Studio term, defined once.
Nexus Studio is an on-prem QA operating system for Playwright, Selenium and pytest that judges tests with 14 verdicts and keeps evidence of every failure.
Each entry below is one literal definition with a link to the page where the term lives. Where a page and this glossary differ, the page is right. Last updated: .
The product
What Nexus Studio is, and where it runs.
What is Nexus Studio?
Nexus Studio is an on-prem QA operating system for Playwright, Selenium and pytest that judges tests with 14 verdicts and keeps evidence of every failure.
What is On-prem QA?
Running tests, keeping evidence and storing history on machines you control. Nexus Studio's only automatic connection is the license check, which carries no test data.
What is Forensic verdict?
One of 14 named outcomes that replace pass or fail. Each tells apart a clean pass, a pass with a reservation, a broken application and an environment problem.
What is an engine host?
The machine where the engine runs your suites: the Windows PC itself, Docker on it, or a Linux server over SSH. The Studio stays on Windows.
The 14 verdicts
Every run gets one of these instead of pass or fail. The decision order is on the verdicts page.
What does SECURE mean?
A run that passed on its own, with no retry and no repaired selector. Nothing to do.
What does CHAOS_RESILIENT mean?
A run that passed while a fault was injected, so the application held. It is filed apart from bug reports, so chaos never skews stability numbers.
What does AUTO_RECOVER mean?
A run that passed only after a broken selector was repaired. The locator in the test is still dead, so it needs fixing.
What does FLAKY mean?
A run that failed and then passed on a retry, with nothing repaired. It signals unexplained instability in the application or the test, not a fix.
What does PERFORMANCE_DEGRADED mean?
A run whose assertions passed but whose page was too slow: a Web Vitals or lazy-loading budget was exceeded. Advisory until you make the check strict.
What does A11Y_VIOLATION mean?
A run whose assertions passed but whose accessibility check did not. The metadata lists each axe-core rule that fired. Advisory until you make the check strict.
What does SECURITY_VIOLATION mean?
A run whose assertions passed but whose security headers did not, for example a missing CSP. Advisory until you make the check strict.
What does BREACHED mean?
A run where an assertion failed and no environment or harness pattern explains the error. It is a real finding to triage as Real bug or False alarm.
What does BREACHED_FLAKY mean?
A run that failed on its last attempt after an earlier retry, so the failure itself is inconsistent. Compare the attempts before filing a bug.
What does CHAOS_BREACHED mean?
A run that broke under an injected fault. It is a resilience finding, kept apart from functional bug reports.
What does TIMEOUT mean?
A run that never answered in time. It is not proof that the application is broken: check how fast the page was before changing any limit.
What does UNREACHABLE mean?
A run where the host could not be reached at all: DNS or the connection failed before any page was observed. It is not a finding about the application.
What does BLOCKED mean?
A run where the site defended itself with a 403, a 429, a block page or a bot challenge. It is not a finding about the application.
What does CRASHED mean?
A run where the browser or process died: a closed page, a DevTools protocol error or a dead WebDriver session. It is infrastructure, not the application.
API tests
REST and GraphQL tests have verdicts of their own.
What are API verdicts?
Seven outcomes for REST and GraphQL tests: API_SECURE, API_FLAKY, API_DRIFTED, API_BREACHED, API_DEGRADED, API_VULNERABLE and API_BLOCKED. Each is judged from the response.
What is Schema drift?
A change in the shape of an API response compared with the last one recorded. It reads API_DRIFTED and counts as a pass until you make it strict.
What does API_BLOCKED mean?
An API run answered with a 429, or a 403 carrying a Cloudflare or Akamai signature, so the request never reached your product.
Evidence and proof
What a run leaves behind, and how it can be checked.
What is Evidence Locker?
The Studio screen where every result is filed. A failure keeps its forensic report, metadata, raw log and screenshot, and a real failure also keeps its video.
What is a sealed chain of custody?
An opt-in ledger that hashes every file of an evidence record with SHA-256 and chains it to the record before, so verification names the first record that changed.
What does UNVERIFIED mean?
The state when no chain of custody ledger exists. Verification reports UNVERIFIED and fails, never a silent OK, because missing proof is not proof.
What is a Real bug or False alarm call?
A triage decision on a failed run, made with a required reason. The call joins the chain of custody.
What is Failure video?
A recording of the session kept for a real failure: BREACHED, BREACHED_FLAKY and CHAOS_BREACHED by default, and TIMEOUT and BLOCKED when you turn them on in Settings.
What is Identity matrix?
A History view of every test against every browser and device, showing where it passes and where it does not. Dotted means it never ran there.
What is Evidence budget?
One total size limit for all stored evidence, 20 GB by default, with a 10 GB free-disk floor. The oldest runs go first; the newest run of each test stays.
What is Flakiness predictor?
A risk score for each test and browser, trained on your own History, that says how sure it is. A single run reads as low confidence.
Creating and repairing tests
How a test gets written, reviewed and kept alive.
What is Nexus Recorder?
Writes Playwright or Selenium code as you click in a real browser, with no AI in the loop, and grades each selector with fixed rules.
What is AI writer?
Drafts tests with the AI tool you already use or a local model. A draft passes the Vet and rehearses in the sandbox; nothing is saved until you accept it.
What is Vet?
A rule-based review of a test: seven checks and no AI, among them a test id, assertions that can fail, and no sleeps, forced clicks, mocks or fragile locators.
What is Test Adapter?
Imports the Playwright or Selenium tests your team already has, rewriting by rule rather than by AI, and blocks the import if an assertion would be lost.
What is Repair line?
A gutter action on a failed line that re-records the step or swaps in a stronger selector, strongest first and XPath last. Saving the change approves it.
What is self-healing that asks first?
After a failure, a background browser replays the test to the failed line and prepares candidate selectors, with no AI. The test stays red until you choose.
What is Locator health?
A list of the weak and medium locators in the test you have open, with the line and the reason, shown before they break.
What is Sandbox rehearsal?
A new test first runs in a sandbox that writes nothing to evidence or history. After three rehearsals it is ready for its first real run.
Running tests
Where and when tests run.
What is Run patrol?
The Studio's run screen. It runs tests on Chromium, Firefox, WebKit and real Android phones, one after another or up to three at once, with one verdict per engine.
What is Watch?
Scheduled re-runs of a selection, from every 5 minutes to every 24 hours or once on a date. It alerts on Slack or Telegram only when a state changes.
How does localhost testing without a tunnel work?
The engine starts your app, waits until it answers, runs the tests and stops it, so a localhost or private staging URL is tested as it is.
What is Self-hosted Playwright?
Running Playwright on machines you control instead of a cloud grid: your PC, a mini PC, Docker or a Linux server over SSH, with the evidence kept on your disks.
Stress, security and reporting
Checks beyond a functional test, and the release call.
What is Chaos Lab?
Injects faults into the browser, such as a slow network, an offline blackout or a CPU throttle, or pauses, stops, kills and restarts a registered Docker container.
What does NOT_INJECTED mean?
The infrastructure chaos outcome when no fault fired. A run where no fault landed is never called resilient.
What is Load?
Worker threads in the engine generate load and report requests per second, p50 to p99 latency, error rate and APDEX, with a verdict of PASS, OVER_BUDGET or UNMEASURED.
What is Smart Monkey?
An autonomous DAST scanner for targets you are authorized to test. It confirms blind findings on a listener on your own machine and exports SARIF.
What is Visual regression?
Pixel comparison against pictures you approved: the first run plants a baseline and later runs are compared with it. Baselines are kept per browser and platform.
What is Executive Report?
A release call of go, caution or no-go, with confidence, coverage and pass rate, plus failures with their root cause, for whoever signs off a release.
Testing concepts
General terms, defined the way this site uses them.
What is Flaky test?
A test that fails and then passes on a retry. Nexus Studio files it as FLAKY rather than as a quiet pass, so an unstable test stays visible.
What is False failure?
A failed test caused by the environment, not the application, such as a slow network, an unreachable host, a bot challenge or a dead browser.
What is Brittle selector?
A locator that breaks when the UI changes. Locator health lists weak selectors before they break, and the Repair line proposes stronger ones.
What is Per-machine pricing?
Plans are priced per activated Windows install, with every feature on every plan and no charge per user, per test or per parallel run.
See the terms on your own app.
Free for 30 days, with every feature, on your machines. No card needed.