Skip to content

xyva.ai — Neural Test Operations Studio

Cloud QA portal with a local Playwright agent. Sign in at xyva.ai/app, connect a one-line agent on your machine, and orchestrate Swarm QA, accessibility audits and Playwright runs from your browser. BYOK for every AI provider.

xyva.ai portal

How it works

  1. Sign in to the portal at xyva.ai/app with Google, GitHub or magic link.
  2. Start the local agent with npx @xyva/agent in the project you want to test.
  3. Trigger runs from the browser — execution, screenshots and logs stay on your machine; the portal stores summaries, billing and team state.

No installer, no code signing, no admin rights. Your repository never leaves your laptop.

Features at a Glance

FeatureAvailabilityDescription
WizardGuided OnboardingSolo + TeamIn-portal setup with project scan, provider config, and preflight checks
Test HubTest CatalogSolo + TeamCentralised test catalog with data-driven parameterisation and dataset management
RunnerPlaywright ExecutionSolo + TeamBrowser-driven runner with parallel execution, debug mode, and CI integration
AI AssistantProject-aware chatSolo + TeamConversational AI with full project context — code generation, reviews, suggestions
Test BuilderTest scaffoldingSolo + TeamGenerate Playwright tests from a URL or description
Swarm QA8 focused agentsSolo + TeamLink Patrol, HTTP Guard, A11y Scout, Smoke Flow, Perf Sentinel, SEO Recon, Form Fuzzer, API Health
Accessibility AuditorBFSG / WCAG 2.1 AASolo + TeamAxe-based audits with evidence reports
Architecture AuditDependency insightTeamProject structure, dependency graph, component analysis
Evidence & ReportsScreenshots & tracesSolo + TeamPDF export, visual regression, structured run reports
Ops PortalInfrastructure healthOn requestPrivate deployment and advanced operations visibility
Swarm ScheduleAutomated scansOn requestRecurring Swarm QA with webhook-based escalation for private rollouts

Swarm QA — 8 Specialised Agents

AgentWhat it does
Link PatrolCrawls all links, checks external domains, detects redirect chains
HTTP GuardMonitors HTTP errors, network failures, console errors
A11y ScoutWCAG/BFSG accessibility checks — alt text, ARIA, contrast, landmarks
Smoke FlowPage-load verification, JS error detection, rendering failures
Perf SentinelCore Web Vitals (LCP, CLS, TTFB), resource sizes, load benchmarks
SEO ReconMeta tags, Open Graph, canonical URLs, structured data, sitemap
Form FuzzerXSS probes, SQL injection, overflow, empty submit, error feedback
API HealthEndpoint availability, response time, status code validation

All agents run in parallel on your local agent. Findings are classified by severity (critical → info) and action category (fix-now → ignore).

Getting Started

Core Features

AI

Compliance & Audit

Private Deployment & Ops

Billing & Teams

Cloud portal · Local Playwright agent.