testgen-ai 0.2.3

The Autonomous QA Agent from Your CLI - AI-powered test generation and execution

🚀 TestGen AI

The Autonomous QA Agent from Your CLI

TestGen AI is a CLI tool that automatically generates, runs, and reports on test suites for your code using LLMs. Point it at a directory, set an API key, and let it handle the rest.

---

📦 Installation

Recommended: Use a virtual environment:

python -m venv testgen-env
# Activate it:
# Windows:
testgen-env\Scripts\activate
# macOS/Linux:
source testgen-env/bin/activate

# Install
pip install testgen-ai

Or install globally:

pip install testgen-ai

Verify:

testgen --version

That's it! Everything is configured automatically. On Windows, the first run will set up your system PATH automatically.

---

⚡ Quick Start

1. Set your API key (one-time)

# Google Gemini (default — has a free tier)
testgen config set GEMINI_API_KEY AIza...

# OpenAI
testgen config set OPENAI_API_KEY sk-...

# Anthropic / Claude
testgen config set ANTHROPIC_API_KEY sk-ant-...

# Ollama (local, no key needed)
testgen config set LLM_PROVIDER ollama

Keys are saved globally to ~/.testgen/.env and apply to every project automatically.

2. Generate tests

testgen generate ./src

3. Run tests

testgen test

4. Generate a report

testgen report

5. Or do everything in one shot

testgen auto ./src

---

✨ Features

• 🤖 AI-Powered Generation — Uses GPT-4, Claude, Gemini, or local Ollama to write real test cases, not boilerplate
• 🌍 14 Languages — Python, JavaScript, TypeScript, Go, Rust, Java, C#, Ruby, PHP, Swift, Kotlin, C++, HTML, CSS
• 👀 Watch Mode — Detects file saves and regenerates tests in real time (--watch)
• 📊 Terminal Dashboard — Color-coded test matrix with pass/fail/skip/duration per test
• 📈 HTML Reports — Professional reports with metrics, coverage insights, and execution distributions
• ⚡ Smart Context — AST-based extraction sends only relevant code to the LLM, keeping token costs low
• 🔄 One-command Workflow — testgen auto runs the full generate → execute → report pipeline

---

🎯 The AGER Architecture

TestGen AI operates on a 4-step loop:

┌──────────┐     ┌──────────┐     ┌──────────┐     ┌──────────┐
│ Analyze  │────▶│ Generate │────▶│ Execute  │────▶│ Report   │
│ (Scanner)│     │ (Brain)  │     │ (Runner) │     │ (Visuals)│
└──────────┘     └──────────┘     └──────────┘     └──────────┘

Phase	What happens
Analyze	Scans your directory, extracts function signatures and docstrings to build minimal context
Generate	Sends context to your LLM and receives executable test code
Execute	Runs the generated tests via the appropriate framework (pytest, Jest, cargo test, etc.)
Report	Renders a live terminal matrix and compiles an HTML report

---

🎨 CLI Commands

Command	What it does
testgen generate <path>	Generate tests for all source files in <path>
testgen generate <path> --watch	Watch mode — regenerate on every save
testgen test	Run all generated tests
testgen test --verbose	Run with full output
testgen report	Build an HTML test report
testgen auto <path>	Full pipeline: generate → test → report
testgen config set KEY VALUE	Save a config value globally
testgen config show	Print current global config

---

📊 Terminal Output

╔══════════════════════════════════════════════════════════════════╗
║                    TEST EXECUTION MATRIX                         ║
╠═══════════════════════════════╦══════════╦══════════╦════════════╣
║ Test Name                     ║ Status   ║ Duration ║ Details    ║
╠═══════════════════════════════╬══════════╬══════════╬════════════╣
║ test_user_login               ║ ✔ PASS   ║ 0.24s    ║            ║
║ test_user_registration        ║ ✔ PASS   ║ 0.31s    ║            ║
║ test_password_validation      ║ ✘ FAIL   ║ 0.12s    ║ AssertionE…║
║ test_database_connection      ║ ✔ PASS   ║ 5.01s    ║ [SLOW]     ║
║ test_api_endpoint_users       ║ ✔ PASS   ║ 0.89s    ║            ║
╚═══════════════════════════════╩══════════╩══════════╩════════════╝

Summary: 4 passed, 1 failed, 0 skipped | Total: 6.57s

---

🌍 Supported Languages

Language	Test Framework Used
Python	pytest
JavaScript	Jest
TypeScript	Jest
Go	go test
Rust	cargo test
Java	JUnit
C#	NUnit / xUnit
Ruby	RSpec
PHP	PHPUnit
Swift	XCTest
Kotlin	JUnit
C++	Google Test
HTML	Custom HTML validator
CSS	Custom CSS linter

The target language's runtime (Node, Go, JDK, etc.) must be installed on your machine. pip install testgen-ai only installs the TestGen AI tool itself.

---

⚙️ Configuration

Config can be set globally (persists across all projects):

testgen config set GEMINI_API_KEY AIza...
testgen config set LLM_MODEL gemini-2.0-flash
testgen config set LLM_PROVIDER gemini

Or per-project via a .env file in the project root (overrides global):

OPENAI_API_KEY=sk-project-key
LLM_PROVIDER=openai
LLM_MODEL=gpt-4o
TEST_OUTPUT_DIR=./tests
MAX_CONTEXT_TOKENS=8000

Supported Providers

Provider	Key name	Default model
Google Gemini	GEMINI_API_KEY	gemini-2.0-flash
OpenAI	OPENAI_API_KEY	gpt-4o
Anthropic	ANTHROPIC_API_KEY	claude-3-5-sonnet
Ollama (local)	(none)	llama3

---

🔧 Optional: Browser / UI Testing

For Playwright-based UI test generation:

pip install testgen-ai[browser]
playwright install

The playwright install step downloads browser binaries and is a one-time setup per machine.

---

🛠️ Technology Stack

Component	Technology
Language	Python 3.10+
CLI Framework	Typer
Terminal UI	Rich
AI Layer	LiteLLM (model-agnostic)
Validation	Pydantic
File Watching	Watchdog
Testing Core	pytest
UI Testing	Playwright (optional)
Reporting	Jinja2

---

🤝 Contributing

1. Fork the repository
2. Create a feature branch: git checkout -b feature/my-feature
3. Make your changes
4. Run tests: pytest
5. Commit: git commit -m "Add my feature"
6. Push: git push origin feature/my-feature
7. Open a Pull Request

---

📚 Documentation

Full documentation: JayPatil165.github.io/TestGen-AI

---

📧 Contact

• Author: Jay Ajitkumar Patil
• Email: [e-mail]
• GitHub: @JayPatil165
• LinkedIn: jay-patil-4ab857326
• Issues: GitHub Issues

---

⭐ Star this repo if TestGen AI saves you time! ⭐
Made with ❤️ by developers, for developers