Quality Engineering

AI QA Test Case Generator

An n8n and OpenAI workflow that turns any requirement into ready-to-use test cases.

Personal Project––– views

Problem: Writing thorough test cases by hand took 30 to 60 minutes per feature, and edge cases and security scenarios still got missed.
Approach: An n8n workflow with two GPT-4o-mini agents (one deterministic for extraction, one for generation), MongoDB conversation memory, validation and error handling, exporting to Google Sheets. It accepts natural language, cURL, Postman, OpenAPI and JIRA formats.
Result: Test cases for a feature in 15 to 25 seconds instead of 30 to 60 minutes, at roughly $2 a month for 1,500 test cases.

Overview

An AI assistant for QA teams. Give it a requirement in almost any format and it returns structured, production-ready test cases in Google Sheets, where the team already works.

What I built

An n8n workflow that accepts six input formats
Two separate AI agents with different temperatures: deterministic for extraction, slightly creative for generation
MongoDB memory so follow-ups like "add security tests" keep context
Validation and error branches at every step, because AI output is probabilistic

Read the full write-up in the blog post.