AI-Powered Test Automation Platform
Manual testing is slow, expensive, and error-prone. QA teams spend hours writing test scripts for every feature change. Small teams often skip testing entirely, leading to production bugs and customer churn.
TestPilot AI uses Claude AI to automatically generate comprehensive test plans from plain-English requirements. It runs browser-based tests in serverless Lambda functions with Playwright, providing AI-analyzed results with actionable insights — no test scripts to write.
Designed a multi-tenant architecture supporting organizations with role-based access
Built AI test plan generation using Claude with structured output parsing
Implemented serverless browser testing using Playwright inside Lambda (custom layer with Chromium)
Created SQS FIFO queue for reliable test job processing with exactly-once delivery
Built real-time progress tracking with DynamoDB streams and polling
Deployed via CloudFront + S3 for global edge caching
Next.js 14 (Static Export) → CloudFront CDN
API Gateway → Lambda (Express.js) for REST API
SQS FIFO Queue for test job queuing
Worker Lambda with Playwright + Chromium layer
DynamoDB (5 tables: Users, Orgs, Members, Projects, Runs)
Claude AI API for test plan generation and result analysis
~12s
Test Generation Time
From requirements to full test plan with AI
$3/mo
Infrastructure Cost
Fully serverless, pay only for actual test runs
< 60s
Test Execution
Full browser-based test suite per run
100%
Zero Servers
Completely serverless — no EC2 instances to manage
I build production-grade serverless applications. Let's discuss your next project.