Introduction

Appev is an autonomous QA platform that "watches" your application to evaluate its quality. It uses Vision-Language Models (VLMs) to navigate your app like a human user, finding logical dead ends, confusing UI, and visual bugs.

Quick Start

Run your first autonomous audit in under 5 minutes.

Configuring Personas

Define traits for the Chaos Engine (Novice, Hacker, etc).

Why Appev?

Traditional testing requires writing endless scripts. User testing is slow and manual. Appev is an "AI User" that navigates your web or mobile app 24/7. It identifies UI bugs, broken flows, and usability friction points.

Zero-setup testing: No brittle scripts. If the UI changes, the AI adapts.
UX Audit: Analyzes flows against established UX principles (Nielsen’s Heuristics).
Visual Intelligence: Understands visual hierarchy, accessibility, and copywriting.

Installation

Get started with Appev by installing our official SDK for Node.js.

Package Manager

Install the core package using your preferred package manager:

Terminal

npm install @appev/sdk

Environment Setup

Add your API key to your environment variables:

.env

APPEV_API_KEY=your_api_key_here

Your First Audit

The fastest way to integrate Appev is via our Node.js SDK. This creates a "Steel Thread"—a single critical path test that runs on every deployment.

tests/critical-path.js

import { Appev } from '@appev/sdk';

const client = new Appev(process.env.APPEV_API_KEY);

// Define the intent in plain English
const session = await client.audit({
  url: 'https://staging.myapp.com',
  intent: 'Log in as a user, add the Pro Plan to cart, and checkout.',
  persona: 'novice_user', 
  viewport: { width: 1280, height: 720 }
});

if (session.score < 85) {
  console.error(`Audit Failed. Score: ${session.score}`);
  console.log(`Video Report: ${session.reportUrl}`);
  process.exit(1);
}

The Chaos Engine

Most testing bots act like robots. Real users act like chaotic humans. Appev doesn't just run a test; it roleplays.

Persona-Driven Chaos

Appev deploys AI agents with specific personalities to stress-test your UX and find business-killing friction points.

The "Boomer" Persona

"Why is this button so small?"

Uses large fonts, clicks slowly, and gets confused by icons without labels.

The "Power User"

"I need shortcuts."

Uses keyboard shortcuts, clicks rapidly, and keeps multiple tabs open.

Frustration Scoring

Appev generates a "Frustration Score." It doesn't just say "The app works." It says: "The functionality works, but the 'Boomer' persona took 45 seconds to find the checkout button."

Semantic Vision

Understanding Intent vs. Error. Traditional visual regression testing is noisy; Appev is intelligent.

Intent vs. Error

Appev uses Vision-Language Models (VLMs) to understand if a change is an intentional design update or a breaking bug.

Design Redesign (PASS)

"The header looks different, but nav links are still present and clickable."

Visual Occlusion (FAIL)

"The 'Submit' button is slides behind a chat widget. A user cannot click this."

Appev Score

Make quality quantifiable. The Appev Score gives you a comprehensive look at your release quality.

When a developer pushes code, Appev generates a "Release Scorecard" (The Appev Score) based on three key pillars:

100%

Functionality

85%

Usability

92%

Performance

Authentication

Secure your audit runs using API keys and secure environment variables.

Authentication details go here...

Webhooks

Set and forget. Schedule tests or trigger them via webhooks.

Webhook configuration details go here...

CI/CD Pipeline

Integrate Appev into your deployment workflow.

CI/CD integration details go here...

JSON Config Schema

Define your personas and test parameters using our structured JSON schema.

appev.config.json

{
  "persona_id": "boomer_novice",
  "name": "Grandpa Joe",
  "traits": {
    "visual_acuity": "low",
    "patience": "low",
    "tech_literacy": "none",
    "input_speed": "slow"
  },
  "goal": "Purchase a gift card for my grandson."
}

Persona List

Choose from our pre-defined personas or create your own custom agents.

boomer_novice
Low visual acuity, low patience, no technical literacy. Simulates older users.
power_user
High speed, keyboard focus, multi-tab usage. Detects race conditions.
hacker_persona
Attempts SQL injection, weird characters in inputs, URL traversal.

Error Codes

Understanding root causes through integrated forensics.

Appev correlates visual failures with browser logs (Network + Console + Vision).

"The checkout UI is frozen because the /api/cart endpoint returned a 500 error."