Runner

The Runner is xyva's execution engine — a neural wrapper around Playwright that adds live streaming, intelligent error analysis, and seamless integration with the rest of the platform.

Execution Modes

Mode	Scope	Use Case
All Tests	Entire test suite	Full regression pass
Surgical	Specific files or tests	Debugging a single spec
Smart Run	Git-affected tests only	Fast feedback on your current changes

Smart Run

Smart Run analyzes git diff to identify which test files — and which source files they depend on — have changed. Only affected tests execute, cutting feedback time dramatically.

Console

The Runner console streams Playwright output in real time with three enhancements:

• Mojibake Protection — detects and fixes character encoding corruption from child processes
• Error Highlighting — failures render in red with clickable source links
• AI Analysis — click the Brain icon next to any error for instant root-cause analysis

Browser Selection

Choose a target browser before launching:

• Chromium — default, matches Chrome and Edge behavior
• Firefox — Gecko engine for compatibility testing
• WebKit — Safari engine for macOS/iOS parity

Preflight Guard

Before every run, the Runner performs a system check:

✓ Node.js v20.11.0 detected
✓ Playwright 1.52.0 installed
✓ Browser binaries: chromium ✓ firefox ✓ webkit ✓
✓ Environment variables loaded (4 variables, 1 sensitive)

Preflight Failure

If any check fails, the run is blocked and the console shows the exact resolution step — for example, npx playwright install webkit to install the missing browser.

Sub-Pages

Page	Content
Execution Modes	All Tests, Surgical, Smart Run explained
Browser Selection	Cross-browser testing with device descriptors
Console	Live output, error highlighting, AI analysis
Debug Mode	Headed browser, step-through, trace viewer
CI Integration	GitLab CI, GitHub Actions, headless mode, reports

Test Hub Integration

Tests launched from the Test Hub arrive in the Runner with their full context — selected scope, bound variables, and dataset rows. The Runner translates this into the appropriate Playwright CLI flags automatically.

Test Hub (scope + data) → Runner (execute) → Intelligence (analyze results)