Voice Agent Tester
A command-line tool that automates testing of voice-enabled web applications using Puppeteer browser automation. It simulates user interactions, manages audio playback and recording, and evaluates agent responses through AI-powered analysis.
Voice Agent Tester provides comprehensive testing capabilities including browser automation, voice interaction support through audio files or text-to-speech synthesis, session recording, audio transcription, performance measurement, and LLM-based response evaluation.
Perfect for quality assurance teams and developers building voice-enabled applications. Execute multiple test iterations, generate detailed performance reports, and ensure your voice agents deliver consistent, high-quality interactions.
Need help running tests or support setting up your testing environment? Reach out to contact@livetok.io.
Browser Automation
Automates testing of voice-enabled web applications using Puppeteer for realistic user interaction simulation.
Voice Interaction Support
Supports audio playback from files or text-to-speech synthesis, recording agent responses, and transcription services.
Performance Measurement
Captures detailed performance metrics including latency, response times, and interaction durations.
AI-Powered Evaluation
Uses LLM-based analysis to evaluate agent responses for accuracy, relevance, and conversational quality.
Session Recording
Records complete test sessions including audio, transcriptions, and browser interactions for analysis.
CSV Reporting
Generates comprehensive CSV reports from multiple test iterations for statistical analysis and trend tracking.
Quick Start
git clone https://github.com/livetok-ai/voice-agent-tester
cd voice-agent-tester
npm install
export OPENAI_API_KEY=your_api_key
npm start -- -a applications/config.yaml -s scenarios/test.yaml