Glossary
Quick reference for the vocabulary used across Testver. If a screen uses a term and you’re not sure what it means, find it here — most entries link to the screen where the term actually appears.
Admin The most privileged role. Can manage users, read the Audit Log, and change org-level Settings — on top of everything a Tester can do.
AI Assistant The built-in chat that helps you write, fix, and analyse tests using a connected LLM. See AI Assistant.
AI Record A tool that captures a browser session (clicks/types/assertions) and turns it into a generated test script. See AI Record.
AI Test Gen Generates test cases from requirements or scripts. See AI Test Gen.
App Mode (Testing Mode) Whether this Testver instance is set up for Web or Mobile automation. Chosen during first-time setup and changeable in Settings → General. Controls which surfaces (Runner, Recorder, Schedules) are shown.
Agent Skill (Skill) A reusable, packaged capability the AI can invoke for a specialised task. See Agent Skills.
API key A secret string from an AI provider (Anthropic, OpenAI, etc.) that lets Testver use AI features. Configured in Settings or via environment variables.
Audit Log The record of who did what, when. Admin-only. See Audit Log.
Automation Coverage
The share of your test cases that have a script mapped (and can thus run automatically): automated test cases / total active test cases. See Coverage.
Catalog The full set of tests Testver discovered by scanning your project (the Test Explorer view). The Dashboard’s health cards are derived from the catalog, not from run history.
Cloud grid (cloud provider) A remote device/browser farm — BrowserStack, LambdaTest, Sauce Labs, etc. — that Testver can run tests on. Configured under Connectors.
Code Map A structural X-ray of your automation codebase — methods, page objects, callers, orphans, and quality signals. See Code Map.
Connector A configured integration with an external service (JIRA, Slack, Email, a cloud grid, etc.). See Connectors.
Coverage What is and isn’t tested. Flavors:
- Source coverage — % of sources with ≥1 linked test case.
- Automation coverage — % of test cases with a script mapped.
Cycle One execution attempt of a test plan. A plan typically has multiple cycles (one per environment or round). See Test Plans.
Defect A bug ticket. Created manually or auto-generated from a failed cycle execution. See Defects.
Default provider
The LLM provider Testver uses unless you pick another at conversation time. Set in Settings or via TESTVER_DEFAULT_PROVIDER.
Evidence Files attached to a manual cycle execution as proof — screenshots, logs, exports. Each cycle execution has its own evidence list.
Execution (Run) One time you executed tests. Produces a result with passed/failed/skipped counts, a duration, and a status. Listed in Results.
Flaky A test that produces different results on identical code — sometimes pass, sometimes fail. The Dashboard surfaces these.
Folder (test case folder) Hierarchical organisation for test cases. Folders nest and counts roll up.
Framework The test framework Testver detected for your project (Playwright, Cypress, Selenium, WebdriverIO, pytest, etc.). Shown in the header.
Healing (self-healing) Testver’s ability to recover from changed locators by recording known-good element context and re-resolving selectors when they drift, reducing flaky failures from minor UI changes.
Incremental Mode (AI Test Gen) A flag telling the AI to skip requirements whose source quote already has matching test cases. Use it after editing a source to cover only the new content.
LLM provider An AI service Testver talks to — cloud (Anthropic, OpenAI, Google, xAI) or local (Ollama, LM Studio). Configured in Settings.
Local Browser Connector A per-user mechanism (token-based) that lets Testver drive a browser on your own machine. See Local Browser Connector.
Mapping (script mapping) The link between a test case (in the catalog) and an automation script (on disk). See Script Mapping.
MCP Model Context Protocol — a standard for extending the AI Assistant with external data sources/tools.
Module A logical grouping of tests, usually a folder. The Dashboard’s Module Distribution shows how tests spread across modules.
Orphan Something that should be linked and isn’t:
- Orphan Test Case — not linked to any requirement/source.
- Orphan Test Script — a mapping whose test case was deleted.
- Orphan Source — a source with no linked test cases.
Output The captured console output of a run (and, for parametrized tests, of each variant). See Output.
Pass rate
The percentage of executed tests that passed in a run: passed ÷ total × 100. Trended on the Dashboard.
Persona An optional attribute on a test case (“Admin”, “Customer”, “Guest”) so coverage can be sliced by persona.
Plan (test plan)
A named scope of test cases for a release/sprint, with an ID like TP-1. Cycles execute against a plan. See Test Plans.
Recording (AI Record) A captured browser session. Its timeline of clicks/types/assertions becomes input for the AI to generate a test script.
Reverse proxy
An HTTPS front-end (nginx, Caddy, Traefik) placed in front of Testver for a team deployment. It must forward WebSocket Upgrade headers so live updates work.
Role A user’s permission level — Admin, Tester, or Viewer. Set on the Users screen.
Run (Execution) See Execution. One execution of tests; the Runner starts them and Results keeps the history.
Schedule A saved combination of “what to run” and “when”. Triggers runs on a clock, with an overlap guard so a slow run never double-fires. See Schedules.
Session One active login on one device. Managed from Profile → Active Sessions; admins can revoke any from Users.
Source Anything that defines what to test — a requirement document, notes, a JIRA ticket, an uploaded file. See Sources.
Strategy (test strategy) A test’s approach — positive, negative, edge case, security, performance, etc.
Suite A test file containing one or more tests. Roughly: suites are files, tests are the cases inside them.
Tag A label attached to a test case. Tags are auto-pruned — when no case uses a tag, it’s deleted.
Tester The day-to-day role: can create, edit, and run tests and use AI features, but cannot manage users or read the Audit Log. The right default for most teammates.
Test Case A documented test — title, preconditions, steps, expected results. Manual or automated. See Test Cases.
Trace (Playwright trace)
A .zip recording of a run including DOM, network, and console. View at trace.playwright.dev.
TC ID / TP ID
Human-readable IDs assigned automatically — test cases are TC-1, TC-2; test plans are TP-1.
Variation / Variant One data set of a parametrized (data-driven) test. Testver rolls the variants of a parametrized test up under a single Variations view, each with its own steps and Output.
Viewer The read-only role: can view cases, plans, results, defects, and dashboards, but cannot run, edit, create, or use token-consuming AI features.
Uncategorized Test cases not assigned to any folder. Always shown at the bottom of the folder list.
Workflow A saved, multi-step automation that chains Testver actions together. See Workflows.