Open-Cluely

Electron desktop meeting assistant with live transcription, screenshots, and Gemini-based assistance.

Project Structure

• src/main.js - main Electron process entry and runtime orchestration
• src/bootstrap/ - startup environment loading, validation, and .env persistence
• src/windows/assistant/ - active Electron window files (window.js, preload.js, renderer.js, renderer.html, styles.css)
• src/windows/legacy/ - old or backup window/transcription experiments kept for reference
• src/services/ai/ - AI service integrations such as Gemini
• src/services/state/ - local persisted state helpers such as cache/app-state.json handling
• src/config.js - source of truth for Gemini and AssemblyAI speech model lists/defaults

Setup

Prerequisites

• Windows 10/11
• nvm-windows 1.2.2 or compatible
• Node.js 20.20.1
• npm 10.8.2
• Gemini API key
• AssemblyAI API key

Native Windows Dependencies

This app uses native Windows modules. If npm ci fails with node-gyp / Visual Studio errors, install:

• Visual Studio 2022 Build Tools or Visual Studio 2022
• Desktop development with C++ workload
• MSVC C++ toolset
• Windows 10/11 SDK
• Python

Install command:

winget install --id Microsoft.VisualStudio.2022.BuildTools --exact --accept-package-agreements --accept-source-agreements --override "--passive --wait --add Microsoft.VisualStudio.Workload.VCTools --includeRecommended"

Install

winget install --id CoreyButler.NVMforWindows --exact --silent --accept-package-agreements --accept-source-agreements
nvm install 20.20.1
nvm use 20.20.1

Restart the terminal or VS Code/Cursor, then run:

node -v
npm -v
npm ci

Run

npm start

For logs:

npm run dev

Configuration

src/config.js

This file is the source of truth for model lists:

• GEMINI_MODELS: Gemini models shown in settings
• ASSEMBLY_AI_SPEECH_MODELS: AssemblyAI speech models shown in settings
• The first item in each list is the default

Current defaults:

• Gemini: gemini-2.5-flash-lite
• AssemblyAI speech: universal-streaming-english

.env

Required:

GEMINI_API_KEY=your_gemini_key
ASSEMBLY_AI_API_KEY=your_assemblyai_key

Optional:

HIDE_FROM_SCREEN_CAPTURE=true
MAX_SCREENSHOTS=50
SCREENSHOT_DELAY=300
NODE_ENV=production
NODE_OPTIONS=--max-old-space-size=4096

Notes:

• GEMINI_MODEL is not read from .env
• available AssemblyAI speech models are not controlled from .env
• HIDE_FROM_SCREEN_CAPTURE=false allows the window to appear in screen share / screenshots

App State

Runtime selections are persisted in:

cache/app-state.json

Stored values:

• selected Gemini model
• selected AssemblyAI speech model

Build

Use:

npm run build -- --config.win.signAndEditExecutable=false

Output:

dist/GoogleChrome.exe

If build fails with a symlink privilege error, enable Windows Developer Mode or run the build from an elevated terminal.

Scripts

• npm start - run the app
• npm run dev - run with logs
• npm run build -- --config.win.signAndEditExecutable=false - build Windows executable

Keyboard Shortcuts

Shortcut	Action
Ctrl+Alt+Shift+V	Toggle voice transcription
Ctrl+Alt+Shift+S	Capture screenshot
Ctrl+Alt+Shift+A	Analyze with AI
Ctrl+Alt+Shift+X	Emergency hide
Ctrl+Alt+Shift+H	Toggle opacity

Note

Use this tool only where recording, transcription, and AI assistance are allowed.