TESTIVA ENGINE V4.0 DOCS

Autonomous QA Documentation

Learn how to deploy Testiva AI's multi-agent swarm into your staging environment or CI/CD pipeline in under 5 minutes. Zero brittle CSS selectors, zero manual test scripts.

5-Min CLI Quickstart

Initialize `@testiva/cli` and trigger your first autonomous DOM exploration run.

GitHub & Vercel CI/CD

Block regressions on every Pull Request and receive automated code patches.

Step 01 • CLI Installation

5-Minute Quickstart

Install the Testiva CLI globally or as a dev dependency in your repository, authenticate with your workspace API key, and point the swarm at any local or staging URL.

# 1. Install the Testiva AI CLI
npm install -D @testiva/cli

# 2. Authenticate with your testiva.xyz workspace
npx testiva auth login --token $TESTIVA_API_KEY

# 3. Dispatch the 5-agent swarm against your staging URL
npx testiva swarm --url https://staging.your-app.com --agents all --auto-pr

Expected CLI Telemetry Output:

✔ Provisioned 5 Firecracker MicroVMs in us-east-1 (boot: 112ms)

➜ [Explorer Agent] Mapped 64 interactive DOM routes & 19 forms

➜ [Hunter Agent] Executed 410 boundary & stochastic fuzzing assertions

✔ Swarm run run_99482a passed (Health Score: 99/100)

Step 02 • Declarative Config

Swarm Configuration (`testiva.config.ts`)

Customize which autonomous agents run, set your concurrency limits, and configure automatic Pull Request remediation. Toggle the switches below to generate your custom configuration file:

Explorer Hunter Fuzz Security OWASP Perf Vitals Auto-PR Fix PII Scrubber

testiva.config.ts

Core Concept • Vision-Language Model

VLM Self-Healing Vision Engine

Unlike Playwright or Cypress, Testiva AI does not bind assertions to brittle CSS classes or XPath selectors. Instead, our Vision-Language Model computes a multi-modal semantic anchor combining visual bounding boxes, accessibility trees, and surrounding layout context.

Legacy Scripted QA (Brittle)

// Breaks whenever Tailwind classes or DOM IDs change
await page.locator('div.checkout-bar > button#btn-pay-v2').click();

Testiva Semantic Intent

// Self-heals across complete UI redesigns
await agent.act("Click the primary payment confirmation button");

Authentication • Stateful Testing

OAuth, 2FA & Session Seeding

To test authenticated dashboards, multi-tenant RBAC roles, and 2FA-protected routes, provide encrypted vault credentials or inject a pre-authenticated `storageState` token header.

auth-personas.yaml

personas:
  - role: "tenant_admin"
    loginUrl: "https://staging.acme.xyz/login"
    email: "${VAULT_ADMIN_EMAIL}"
    password: "${VAULT_ADMIN_PASSWORD}"
    totpSecret: "${VAULT_ADMIN_TOTP_SEED}" # Auto-generates 6-digit 2FA codes

  - role: "readonly_member"
    # Used by Security Agent to verify zero IDOR privilege escalation
    injectHeaders:
      Authorization: "Bearer ${VAULT_READONLY_JWT}"

Autonomous Fixes • GitHub App

Auto-PR Code Remediation

When the Hunter or Security agent reproduces a failure twice (eliminating false positives), our Remediation Engine correlates the browser stack trace with your repository AST and opens a side-branch Pull Request containing the minimal verified fix.

Zero Unreviewed Merges: Testiva's GitHub App only pushes commits to isolated `testiva/fix-*` branches and never merges into `main` without human approval.

Step 03 • Continuous Deployment Gate

CI/CD Pipeline Integration

Trigger Testiva AI automatically on every Pull Request preview deployment.

name: Testiva Autonomous QA Swarm
on: [pull_request]

jobs:
  autonomous-qa:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - name: Run Testiva AI Swarm against Preview URL
        uses: testiva-ai/swarm-action@v4
        with:
          api-key: ${{ secrets.TESTIVA_API_KEY }}
          target-url: ${{ steps.deploy.outputs.preview_url }}
          agents: "explorer,hunter,security,perf"
          fail-on-severity: "high"
          auto-open-pr: true

REST API v1 • Interactive Console

Trigger Swarm via REST API

Dispatch an autonomous test run programmatically from any custom deployment orchestrator. Click Send Test Request below to simulate a live API call:

POST https://api.testiva.xyz/v1/swarm/dispatch

Request Body (JSON)

{
  "target_url": "https://staging.acme.xyz",
  "concurrency": 15,
  "agents": ["explorer", "hunter", "security"],
  "webhook_url": "https://acme.xyz/hooks/qa"
}

Response Payload Ready

// Click "Send Test Request" to inspect live response...

Enterprise Enclave • Private Subnets

Self-Hosted Kubernetes & VPC Runners

For air-gapped staging environments, deploy the stateless Testiva Runner Helm chart directly inside your AWS, GCP, or Azure Kubernetes cluster.

Helm Deployment

helm repo add testiva https://charts.testiva.xyz
helm install testiva-runner testiva/swarm-enclave \
  --namespace testiva-system --create-namespace \
  --set runner.enclaveToken=$TESTIVA_ENCLAVE_TOKEN \
  --set runner.zeroDataRetention=true