Kidari Studio's Automated QA Test Case Generation — A Use Case of Amazon Bedrock and LangGraph | Amazon Web Services

KidariStudio built an automated QA test case generation system with Amazon Bedrock’s Claude models and LangGraph, cutting writing time dramatically.

The team evolved a linear LangChain PoC into a multi-agent LangGraph workflow with input-based routing, role separation, and Human-in-the-Loop review.

It now generates QA test case drafts from production URLs, images, and planning docs in 2–5 minutes, covering about 70–80% of cases.

Model-specific placement and state management improved quality, reduced hallucinations, and eased the release bottleneck for the webtoon/web novel platform.