Qarma
Team consisting of an Ericsson Intern and DevAlly Software Engineer, skilled in full-stack Next.js, Python, and Gemini AI integration.
Project Description
Core Functionality:
Qarma is an AI-powered web testing platform where users speak or type natural language goals, and autonomous agents
navigate websites to validate functionality, returning voice-summarised results.
Stability & Complexity:
Functional prototype with authentication, test creation, execution tracking. Orchestrates voice
transcription, browser automation, LLM reasoning, cloud storage, and audio synthesis in a cohesive workflow.
Innovation:
Voice-first test creation eliminates coding barriers. Multi-agent orchestration executes complex user journeys. AI
summarisation with audio playback makes results accessible to non-technical stakeholders. Natural language replaces
traditional test scripting.
Real-World Impact:
Democratises QA for product managers, designers, and non-developers. Accelerates regression testing. Enables
voice-driven accessibility testing workflows.
Theme Alignment:
- Browsers: Autonomous navigation agents execute tests
- Voices: ElevenLabs speech-to-text/text-to-speech for input/output
- Clouds: Firebase Functions orchestrate distributed test execution
- Tools: Browser automation + LLMs become intelligent testing agents
Technologies:
Next.js 16, TypeScript, React, Firebase (Firestore/Functions/Auth), Clerk, ElevenLabs API, Tailwind CSS, Lucide
Icons, browser automation framework (implied), LLM integration for goal evaluation and summarization.