https://github.com/innovatorved/realtime-interview-copilot
Realtime Interview Copilot is a web application that assists users in crafting responses during interviews. It leverages real-time audio transcription and AI-powered response generation to provide relevant and concise answers.
https://github.com/innovatorved/realtime-interview-copilot
ai chatgpt innovatorved interview interview-preparation llm realtime speech-to-text
Last synced: about 1 month ago
JSON representation
Realtime Interview Copilot is a web application that assists users in crafting responses during interviews. It leverages real-time audio transcription and AI-powered response generation to provide relevant and concise answers.
- Host: GitHub
- URL: https://github.com/innovatorved/realtime-interview-copilot
- Owner: innovatorved
- License: other
- Created: 2024-02-28T19:24:32.000Z (over 2 years ago)
- Default Branch: main
- Last Pushed: 2026-05-03T12:13:19.000Z (about 1 month ago)
- Last Synced: 2026-05-03T14:15:22.158Z (about 1 month ago)
- Topics: ai, chatgpt, innovatorved, interview, interview-preparation, llm, realtime, speech-to-text
- Language: TypeScript
- Homepage:
- Size: 1.14 MB
- Stars: 89
- Watchers: 3
- Forks: 18
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- Contributing: CONTRIBUTING.md
- License: LICENSE
Awesome Lists containing this project
README
# Realtime Interview Copilot
Real-time AI copilot for interviews. Captures your system audio, transcribes it live with Deepgram, and answers questions on the fly using multimodal AI (text + screenshots). Desktop app window is hidden from screen shares.
## Download
Get the latest build from the [**Releases page**](https://github.com/innovatorved/realtime-interview-copilot/releases/latest) — macOS (Apple Silicon) `.dmg` or Windows (x64) `.exe`.
### macOS
```bash
brew tap innovatorved/tap
brew install --cask realtime-interview-copilot
```
Or grab the DMG and right-click the app → **Open** on first launch.
### Windows
Run the installer. If SmartScreen warns, click **More info → Run anyway**.
## Features
- 🎙️ **Live transcription** of system audio (Deepgram) — no BlackHole / VB-Cable needed
- 🤖 **Ask AI** with streaming answers (Gemini / OpenAI-compatible)
- 🖼️ **Vision** — press `⌘⇧1` to snap your screen and ask AI about it
- 🕵️ **Undetectable** — window is hidden from screen shares and full-screen spaces
- ⌨️ **Shortcuts** — `K` focus Ask AI, `C` Copilot mode, `S` Summarizer, `Escape` clear
## System Audio — No Virtual Drivers
The app uses Electron's native display-media loopback.
- **macOS**: grant *Screen Recording* once (System Settings → Privacy & Security), relaunch. Silent after that.
- **Windows**: works out of the box (WASAPI loopback).
- **Linux**: use any PulseAudio / PipeWire monitor source.
## Development
```bash
git clone https://github.com/innovatorved/realtime-interview-copilot.git
cd realtime-interview-copilot
bun install
bun run electron:dev # dev
bun run electron:build # package installers into dist/
```
Requires Node 20+ and [Bun](https://bun.sh) 1.3+.
## Stack
React · Next.js · Electron 41 · Tailwind · Shadcn/UI · Deepgram · Gemini · Cloudflare Workers + D1
## Releases
Tagged pushes (`v*`) trigger a GitHub Actions workflow that builds macOS and Windows artifacts and publishes them to a GitHub Release automatically.
## Contributing
PRs welcome. See [CONTRIBUTING.md](./CONTRIBUTING.md). Windows/Linux test reports especially appreciated.
## Architecture
For a complete walkthrough — audio capture, transcription pipeline, AI/vision backend, worker API, CI, distribution, and security — see [ARCHITECTURE.md](./ARCHITECTURE.md).
## License, Privacy & Contact
[LICENSE](./LICENSE) · [Privacy Policy](./PRIVACY.md) · [Architecture](./ARCHITECTURE.md) · [vedgupta@protonmail.com](mailto:vedgupta@protonmail.com) · [Report issues](https://github.com/innovatorved/realtime-interview-copilot/issues)
---
⚠️ For educational use. Check your interview platform's terms before using any AI assistance tool.