mobile-tester-agent
Introduction: AI-powered test automation tool which allows developers to run mobile automated tests
Tags:
The Mobile Tester Agent is an AI-powered test automation tool built with Koog. It allows developers to run automated tests on Android apps using LLM-based agents.
📂 Repositories
- Core Agent (Koog + Ktor API): mobile-tester-agent
- Frontend Dashboard: mobile-tester-agent-frontend
- Sample Android App: mobile-tester-agent-sample-app
🎯 Project Goal
- Automate Android app testing using AI.
- Interpret natural language test scenarios with LLMs (Gemini, GPT, LLaMA, etc.).
- Execute tests on an Android Emulator or Physical Device via ADB.
- Generate reports for developers and QA teams.
- Provide a web dashboard for configuring tests and visualizing results.
🏗️ Architecture
graph TD
A[Frontend Dashboard] --> B[Ktor API]
B --> C[Koog Agent]
C --> D[ADB Tools]
D --> E[Android Emulator / Physical Device]
C --> F[LLM Engine]
C --> G[Test Reports]
G --> A[Frontend Dashboard]
🔄 Workflow
- User Input (Frontend): Define goals, steps, and agent settings.
- Ktor API: Sends scenarios to the Koog Agent.
Koog Agent:
- Uses an LLM to parse steps.
- Executes actions via ADB.
- Device Execution: Runs tests on emulator/physical device.
- Report Generation: Results are returned to the frontend dashboard.
🖥️ Components
- Frontend Dashboard → Create test scenarios, configure AI, view reports.
- Ktor API → Orchestrates communication between frontend and agent.
- Koog AI Agent → Uses LLMs for reasoning, planning, and execution.
- ADB Tools → Executes actions on the device.
- Android Device/Emulator → Runs the target app under test.
📊 Example Test Scenario
Goal: Test the login functionality.
Steps:
- Open app
- Enter username/password
- Click login
- Verify success message
The agent interprets this scenario, executes steps via ADB, collects logs/screenshots, and generates a report.
🚀 Future Improvements
- Accessibility testing.
- Parallel multi-device testing.
- iOS support.