Vectra QA Documentation¶

Welcome to the Vectra QA documentation. Vectra QA is a multi-agent autonomous testing framework that deploys specialized AI agents to explore, test, and validate web applications.

What is Vectra QA?¶

Traditional E2E testing relies on static scripts and brittle selectors. Vectra QA treats testing as an autonomous exploration problem:

🤖 Dynamic Agent Spawning: Orchestrator instantiates UI Explorers and Data Validators on-demand
🧠 Obsidian Memory Layer: Agents read/write Markdown files with YAML frontmatter for structured state tracking
📡 Live Command Center: Dark-mode dashboard with Server-Sent Events for real-time monitoring
💬 Natural Language Interface: Chat with Vectra to configure and run tests conversationally

Quick Links¶

Getting Started — Install and configure Vectra QA
Quickstart — Run your first test in 5 minutes
Architecture — Understand how agents communicate and share memory
API Reference — REST API documentation for all endpoints
User Guide — Learn to write effective test scenarios

Test Types¶

Vectra QA supports multiple test types, each handled by specialized agents:

Test Type	Agent	Description
Homepage	UI Explorer	Page structure, navigation, CTAs, footer, console errors
Navigation	UI Explorer	Link validation, page transitions, mobile menu
Contact Form	UI Explorer	Form fields, validation, accessibility
API Monitoring	Data Validator	Backend API calls, response validation
Accessibility	UI Explorer	WCAG compliance, ARIA, keyboard navigation, alt text
Responsive Design	UI Explorer	Multi-viewport testing (desktop, tablet, mobile)
Full Suite	UI Explorer	Complete audit — all test types combined

System Architecture¶

graph TD
    A[Command Center] -- HTTP/MCP --> B[MCP Server]
    B -- Spawn --> C[UI Explorer Agent]
    B -- Spawn --> D[Data Validator Agent]
    C -- Read/Write --> E[Obsidian Vault]
    D -- Read/Write --> E
    A -- SSE --> F[Dashboard UI]

Getting Help¶

License¶

Vectra QA is released under the MIT License.