Concepts

The terms Arcus uses, grouped by where you meet them.

Structure

Project sits at the top of the hierarchy and holds the test cases, runs, and plans for one application, along with its own settings.

Folder groups test cases within a project, by feature, functionality, or test type. Folders nest, so a hierarchy can match how your application is built.

Module is the part of the application a test case covers. Coverage is reported per module.

Authoring

Test case is a list of steps with expected results, and the building block of your library.

Test step is one action inside a test case, with the result you expect from it.

Step group is a saved sequence of steps reused across test cases. Write a login flow once, and link it wherever it is needed.

Data set holds reusable input values, so one test case runs against several sets of data without being rewritten.

Test case properties are the 4 built-in attributes every test case carries: Test Case Priority, Test Case Status, Test Case Type, and Automation Type. Their values are renamed under Settings > Manage Properties, and new ones cannot be added.

Custom field adds project-specific information to a test case, and becomes a filter on the library.

Label organizes test cases and test runs for filtering and tracking.

Test Capture is a browser extension that records screen, video, and network logs alongside manual testing, and turns them into test cases or bug reports.

Version records a change to a test case. Versions compare against each other, so you can see what changed and when.

Review Management puts a test case through approval before it counts as ready. A reviewer accepts it or sends it back.

Test case template is a starting shape for a new test case, so a team writes them in a consistent form.

Run history shows every execution of one test case, across every run it appeared in.

Generation

QI Home is where test cases are generated, coverage is tracked, and automation happens.

Context source is what generation reads from: a Jira, Azure DevOps, Linear, or ClickUp story, a Confluence page, a Figma design, a file such as a PDF or Word document, or a video recording. Each needs its integration connected.

Ad-Hoc generates on demand. You attach context sources yourself, write a prompt, and start generation.

Sprints generates automatically. When a sprint starts in your connected project management tool, you confirm, and generation pulls the context linked to the sprint stories.

Session is one generation run, from the context you attached to the test cases it produced.

Prompt describes what you want covered. It shapes what generation produces from the same context.

Scenario groups generated test cases by feature area or by the story they came from. Review works scenario by scenario.

Generated test case is a proposal. It enters your library when you accept it.

Gap is a pending test case nobody has accepted, and the signal that an area is not yet covered.

Coverage and quality

Coverage is the percentage of generated test output you have accepted, calculated as accepted divided by accepted plus pending. Rejected cases are excluded. Coverage and Gaps breaks the same figure down per module.

Quality Intelligence Metrics are the 5 measures in the session header, shown in evaluation order because each builds on the one before it.

• Coverage asks whether enough has been tested
• Pass Rate asks whether what was tested is working
• Confidence asks whether those results can be trusted
• Release Readiness combines the 3 into one score
• Release Gate turns that score into a decision

Test plan sets the scope, goals, and schedule for a round of testing, and links test cases to test runs. Plans are generated in 4 types, each a superset of the one before it: Smoke, Feature, Regression, and Deep Regression.

Test run records execution: which test cases ran, what each one returned, and any defects logged.

Requirement traceability report shows which test cases cover which requirements, and where coverage is missing.

Automation

Agentic Learning walks a test case through your live application, observes how it behaves, learns the element interactions, and generates automated steps.

Copilot executes automated steps in a browser, with a live panel showing each step as it runs. It pauses, takes new steps mid-run, and saves the result.

Testsigma Lab is the cloud where a learning or execution session runs. The alternative is your own local device.

Automated test case has steps that execute without anyone performing them. A test case is manual until Agentic Learning automates it.

Arcus CLI runs tests from your own machine, outside the browser.

Developer context

Developer Context Mapping keeps coverage aligned with what the team is building, by reading what developers do in their own tools.

Unlinked Contexts is the Context Management tab holding activity detected from Claude Code, Cursor, GitHub Copilot, Codex, and GitHub pull requests, before anyone has assigned it to a module. Once mapped, an entry moves to Linked Contexts.

Mapping assigns a context entry to a module and sprint. Each entry arrives with a suggestion, and mapping it triggers test case generation or an update.

Connecting other tools

Integration connects Arcus to a tool your team already uses, for requirements, defects, context, or notifications.

Two-way integration syncs in both directions, so test cases and runs are managed from Jira or Azure DevOps as well as from Arcus.

API key authenticates a call to Arcus from outside the application, such as a CI pipeline or a script.

Import brings an existing library into Arcus, from a CSV file or straight from another tool with Quick import. Export takes yours out.