CoverIt is built as a group of services that communicate through the API, queues, shared contracts, and graph storage. The frontend is the main user surface, while the workers do the long-running automation work in the background.
- The QA engineer creates a project, application, and application version in the frontend.
- The API stores project data in Postgres and sends background jobs through Redis.
- The crawler worker opens the target application in Playwright, explores clickable elements and input fields, and records the discovered states and transitions.
- Neo4j stores the state graph, including pages, transitions, locators, and graph relationships.
- DocGen labels states and transitions and generates human-readable artifacts such as guides, BDD-style scenarios, and videos.
- The regression generator converts reviewed flows into a Playwright regression codebase.
- Regression runs send status, scenario results, artifacts, and integration reports back to the API.
- The frontend shows crawl progress, generated flows, run results, project activity, and dashboard analytics.
| Store | Used for |
|---|---|
| Postgres | Users, projects, applications, versions, crawl sessions, test flows, regression runs, artifacts, integrations, and dashboard records. |
| Neo4j | Crawled application graph: states, transitions, graph IDs, semantic labels, and relationships. |
| Redis | Background job queues, worker coordination, manual recording tickets, and asynchronous processing. |
Every crawled state is stored as a node connected to the states it can reach, the action that caused each transition, and the application version it belongs to. That connectivity is what lets later modules generate flows that cover real user behavior, rather than simply confirming that individual elements exist on a page.
| Worker | Responsibility |
|---|---|
| Crawler | Drives the browser, discovers states, records transitions, handles automatic crawl sessions, and supports manual recording. |
| DocGen | Labels graph data, describes pages and actions, waits for label completion, and creates documentation/video artifacts. |
| Regression generator | Reads reviewed flows and emits executable TypeScript/Playwright regression files. |
| Regression framework | Runs generated tests, evaluates assertions, stores artifacts, and reports results back to CoverIt. |