TestGen AI

An AI-powered universal QA automation agent that generates professional Playwright test cases, POM structure, test data, and Jira tickets automatically from one line of description.

Built by Ahtsham Ijaz — QA Automation Lead & AI QA Consultant.

---

✨ Key Features

• 🌐 Universal — Works on ANY web application
• 🔐 Dynamic Credentials — Enter your own app credentials at runtime
• 🎭 POM Structure — Generates Page Object Model automatically
• 📊 Data Driven — Separate JSON test data file
• 🎫 Jira Integration — Auto creates tickets with ADF format
• 📄 Documentation — Generates markdown report automatically
• 🤖 AI Powered — Claude Sonnet API for intelligent generation

---

What It Does

1. Takes a feature description as input
2. Asks for app URL and test credentials
3. Uses Claude AI to generate structured test cases
4. Generates Page Object Model Playwright tests
5. Creates separate JSON test data file
6. Automatically creates Jira tickets for each test case
7. Exports full markdown documentation report

---

🚀 What Gets Generated

From one line input: python3 agent.py "user can login"

Agent automatically creates:

Output	Description
Test Cases	Positive, negative, edge cases
pages/LoginPage.js	POM class with generic selectors
pages/InventoryPage.js	POM class for post-login page
tests/login.spec.js	Data driven Playwright spec
test-data/test_data.json	Separate JSON test data
Jira Tickets	One ticket per test case
qa_report.md	Full documentation report

---

📊 Time Savings

Task	Manual Time	With Agent
Write test cases	2-3 hours	30 seconds
Write Playwright tests	2-3 hours	30 seconds
Create Jira tickets	45 mins	30 seconds
Write documentation	45 mins	30 seconds
Total	6-7 hours	~2 minutes

---

Tech Stack

Layer	Tool
AI Model	Anthropic Claude Sonnet API
Test Framework	Playwright (JavaScript)
Project Management	Jira Cloud REST API
Language	Python 3.9+
Test Data	JSON data driven approach
Architecture	Page Object Model (POM)

---

Project Structure

testgen-ai/ ├── agent.py # Main entry point ├── tools/ │ ├── test_generator.py # Claude AI test case generator │ ├── playwright_gen.py # POM Playwright code generator │ ├── data_generator.py # JSON test data generator │ ├── jira_tool.py # Jira REST API integration │ └── markdown_export.py # Documentation generator ├── generated/ # Auto-generated output │ ├── pages/ # POM page classes │ ├── tests/ # Playwright spec files │ └── test-data/ # JSON test data files ├── .env # API keys (never committed) ├── .env.example # Safe template for others ├── .gitignore └── README.md

---

Setup

1. Clone the repo

git clone https://github.com/Ahtsham-QA/qa-copilot-agent.git
cd qa-copilot-agent

2. Create virtual environment

python3 -m venv venv
source venv/bin/activate

3. Install dependencies

pip install -r requirements.txt

4. Configure environment variables

cp .env.example .env

Then open .env and fill in your actual API keys.

---

Environment Variables

Variable	Description
ANTHROPIC_API_KEY	Get from console.anthropic.com
JIRA_BASE_URL	Your Jira domain e.g. https://yourname.atlassian.net
JIRA_EMAIL	Email you use to log into Jira
JIRA_API_TOKEN	Get from id.atlassian.com/manage-profile/security
JIRA_PROJECT_KEY	Your Jira project key e.g. SCRUM

---

Usage

Interactive mode

python3 agent.py

Inline mode

python3 agent.py "user can login with valid credentials"

Agent will ask for:

🌐 App URL: https://your-app.com ✅ Valid username: your_username ✅ Valid password: your_password ❌ Invalid username: wrong_user ❌ Invalid password: wrong_pass 🔒 Locked username: locked_user

---

🌐 Universal App Support

Agent works on ANY web application:

SauceDemo (username based):

python3 agent.py "user can login"
# URL: https://www.saucedemo.com
# Username: standard_user
# Password: secret_sauce

Automation Exercise (email based):

python3 agent.py "user can login"
# URL: https://automationexercise.com
# Username: your@email.com
# Password: yourpassword

Same agent. Different apps. Everything generated automatically.

---

Skills Demonstrated

• AI agent architecture and prompt engineering
• Anthropic Claude API integration
• Universal Page Object Model generation
• Data driven testing with JSON
• Jira REST API with ADF (Atlassian Document Format)
• Python project structure and modular design
• Real-world QA automation thinking
• CI-ready modular code structure

---

⚠️ Proprietary Software

This software is proprietary and confidential. Viewing is permitted but copying, modifying, or distributing requires written permission.

For licensing inquiries contact: LinkedIn

---

👤 Author

Ahtsham Ijaz — QA Automation Lead & AI QA Consultant

• 7+ years QA experience | 5 years lead roles
• Specializing in AI-assisted test automation
• Available for QA consulting projects

🔗 LinkedIn 🐙 GitHub 📧 Open to consulting inquiries