DEV Community

Cover image for ๐Ÿš€ Atomiq TestDroid โ€” QA Meets Agentic AI ๐Ÿค–
vikas padiyar
vikas padiyar

Posted on

๐Ÿš€ Atomiq TestDroid โ€” QA Meets Agentic AI ๐Ÿค–

Excited to share a glimpse of the Atomiq TestDroid architecture and UI weโ€™ve been building.
๐Ÿ”น What is it?
Atomiq TestDroid analyzes your GitHub repo and generates Java unit tests or a full BDD framework using CrewAI-powered agents.
๐Ÿ”น How does it work?
The architecture is designed around 4 layers:
1๏ธโƒฃ Agentic Core โ€“ Specialized AI agents (Code Cartographer, Unit Test Writer, BDD Author, Planner) analyze the repo, map risks, and auto-generate high-value test cases.
2๏ธโƒฃ Test Harness Layer โ€“ Spins up embedded mock REST services, H2 databases, Mongo containers, EMS/Kafka stubs for realistic component-level validation.
3๏ธโƒฃ Generation & Customization โ€“ Developers can input prompts, prioritize modules, upload requirement docs, and choose between unit test suites or BDD frameworks.
4๏ธโƒฃ Services Layer (FastAPI) โ€“ All agent requests, repo analysis tasks, and test generation pipelines are orchestrated through FastAPI microservices, making the system lightweight, modular, and cloud-ready.
๐Ÿ”น The UI (Angular)
As simple as pasting your GitHub URL, choosing your target (Unit tests or BDD), and clicking Generate. The backend agents take care of the rest.

โœจ Why it matters
With Atomiq TestDroid, QA shifts from being reactive to proactively intelligentโ€”agents decide what to test, how to test, and why it matters.
๐Ÿ‘‰ This is a step toward making test automation self-adaptive, risk-driven, and developer-friendly.

๐Ÿ”— Repo reference: https://github.com/vikaspad/atomiq-droid-suite

Top comments (0)