🤖 AI Interview Assistant - Sell Project

Selling full source code rights to a stealthy, real-time AI interview co-pilot built with Electron, React, Vite, and TypeScript. The app runs as an undetectable, always-on-top desktop overlay that transcribes interview calls locally and streams structured answers using free LLM APIs. 🔥 KEY FEATURES: • Absolute Stealth Overlay: Uses native OS display exclusion (WDAEXCLUDEFROMCAPTURE on Windows, kCGWindowSharingNone on macOS) to stay invisible during screen sharing on Zoom, Google Meet, Teams, OBS, and web browsers. • Dual Audio Stream + Silero VAD: Captures both system audio (interviewer) and microphone (candidate). Automatically fires answer generation after 600ms of interviewer silence. • Local & Secure STT: Built-in local Whisper.cpp integration for fast, offline voice transcription. • Zero-Cost Multi-Key Rate-Limit Fallback: Automatic mid-stream fallback across free OpenRouter models (MiniMax 2.7, Gemma 4, Nemotron 3) with multi-key rotation to guarantee zero downtime. • Local RAG Engine: Ranks and chunks candidate resume + job description via TF-IDF cosine similarity to inject tailored context into every response. • Shortcut Mechanics: Global Ctrl/Cmd+Shift+X hotkey to toggle mouse click-through for zero interaction friction. 📦 WHAT IS INCLUDED: • Complete, clean GitHub repository (Electron + React + Vite + TypeScript) • Full intellectual property and transfer rights • Complete setup instructions and environment scripts 💡 BUYER OPPORTUNITY: This codebase is 100% complete and ready to be packaged into a paid desktop utility, sold as a developer boilerplate, or launched as a SaaS product.