Skip to content

Quick Start

Get from zero to your first test result in under 5 minutes.

Step 1: Install xyva

Download the installer for your platform from xyva.ai and run it. See the full Installation guide for details.

Step 2: Launch and Complete the Wizard

On first launch, the Wizard appears. Walk through these screens:

  1. Language — Select English or German
  2. Auth Doctor — Connect an AI provider or choose Observer Mode
  3. Project Import — Open an existing Playwright project or scaffold a new one
  4. Preflight — Validate Node.js, Git, and browser binaries

TIP

If you do not have an AI API key yet, choose Observer Mode. You can still run tests and use all non-AI features.

Step 3: Import a Project

Point xyva at an existing Playwright project directory. The app automatically detects your framework, counts page objects, fixtures, and test files, and maps the project architecture.

bash
# Example: if you have an existing Playwright project
cd ~/projects/my-app
ls playwright.config.ts  # xyva looks for this file

If you do not have a project yet, use Initialize New Operation to scaffold one:

bash
# xyva runs this under the hood
npm init playwright@latest

Step 4: Run Your First Test

  1. Navigate to the Runner panel in the sidebar
  2. Your test files appear in the file tree
  3. Click the Play button next to any test file or individual test
  4. Watch results stream in real-time in the results panel

You can also run tests from the integrated terminal:

bash
npx playwright test --reporter=list

Step 5: Explore AI Features

With an AI provider connected, unlock the full xyva experience:

FeatureWhat It Does
AI AssistantAI chat for test generation, debugging, and code explanations
Swarm QAMulti-agent scanning (a11y, links, performance, SEO, …)
Project ScoutExploration-based suggestions based on coverage gaps
Auto-Fix suggestionsAssistive selector/fix proposals when tests break — you review and accept

Launch a Swarm Scan

  1. Open the Swarm QA panel
  2. Enter a target URL (e.g., https://your-app.dev)
  3. Select agents: Accessibility Scout, Link Patrol, HTTP Guard
  4. Click Launch Swarm
  5. Review findings in the results dashboard

WARNING

Swarm scans make real HTTP requests to the target URL. Only scan applications you have permission to test.

What Next?

Local-first QA orchestration.