Power Interview is a privacy-first AI assistant designed to help you ace technical and behavioral interviews. With real-time transcription and intelligent suggestions, you'll have the confidence and support you need during live interviews - all while keeping you in control of your data.
Power Interview is built with privacy as a core principle. Here is exactly what is stored where:
- Your interview configuration is stored on your account: full name, CV/profile, and context are saved to the backend so they follow you across devices. You can change or clear them at any time from the configuration dialog
- Transcripts are never stored: interview audio is relayed through the backend to the ASR provider for live transcription and is not retained; transcripts exist only in memory for the duration of the session
- Device settings stay on the device: audio device, window bounds, and scroll preferences are local-only
- Local storage: credentials and local settings are written by Electron Store to your OS user-data directory as a plain JSON file, protected by filesystem permissions
- No Data Mining: no selling or sharing of personal data
- Minimal Data Transfer: only the data needed for an active AI request, or an explicit configuration update, is sent to the backend
Stay on top of the conversation with live ASR:
- Dual-channel transcription (you + interviewer)
- WebSocket streaming for low latency
- Speaker detection
- Full transcript history
- Personalized responses based on CV and job description
- Streaming responses in real time
- Context-aware outputs
- Natural language responses
- Screenshot-based problem understanding
- Multi-image support (up to 4)
- LLM-powered solutions
- Syntax-highlighted code output
- Profile management (CV, job description, etc.), synced to your account across devices
- Audio device selection (local to each device)
- Language support (English)
- Persistent settings
Power Interview desktop client is supported on:
- Windows 10/11 (x64 installer build)
- macOS (Apple Silicon and Intel release artifacts)
Release binaries are published on the GitHub Releases page.
Power Interview follows a client-server architecture.
- Electron + React + TypeScript
- Zustand + React Query
- Handles UI and orchestration
- AI/LLM Service for suggestions
- ASR Service for transcription
- Auth Service
- WebSocket (real-time)
- REST API
- Node.js v22.15+ (required by the main-process test runner)
- pnpm (the repo pins
pnpm@11.1.3viapackageManager)
git clone https://github.com/PowerInterviewAI/client-app
cd client-app
pnpm installpnpm startpnpm lint # ESLint
pnpm build # tsc + vite build (renderer)
pnpm test:main # main-process checks- Set profile (CV, job description)
- Select microphone
- Start assistant
- Code suggestions
- Debugging assistance
- Live transcription
- AI-generated responses
- Context-aware answers
- Self-monitoring
- Feedback loops
- HTTPS + secure WebSockets for all backend traffic
- Interview configuration stored against your account; transcripts never persisted server-side
- Local settings and credentials stored in your OS user-data directory
- Full user control - configuration can be edited or cleared at any time
- Electron
- React 19
- TypeScript
- Tailwind CSS
- FastAPI (Python 3.12), MongoDB, Redis
- REST over HTTPS and WebSocket streaming for ASR
power-interview-client/
├── src/
│ ├── main/ Electron main process
│ └── renderer/ React renderer
├── test/ Main-process checks (pnpm test:main)
├── public/
├── build/
Pull requests welcome.
MIT License
- Email: team@vectorleappulse.xyz
- GitHub Issues for bugs/features
Built to help you succeed in interviews