Learn Glass Cloak

Download the App

Download Glass Cloak for your platform from the website. The app comes as a lightweight .dmg for macOS or .exe for Windows.

Enter Your Access Key

On first launch, you'll see the activation screen. Enter your access key to unlock the app. You can find your key in your dashboard after purchasing.

Download AI Models

Glass Cloak runs AI locally on your machine. Head to Settings → Model Manager to download the required models. This is a one-time setup.

You're All Set!

Glass Cloak is ready to use. The floating bar will appear on your screen. You can toggle it with the keyboard shortcut or the dock icon.

The expanded floating bar (600×54px) shows the full interface: input field, audio controls, and action buttons. This is where you type prompts and manage your session.

• ● Text input for typing prompts directly
• ● Mic button to start voice capture
• ● Screenshot & settings quick-access buttons

Click the logo or press the shortcut to collapse the bar into a tiny pill. It stays accessible but nearly invisible, perfect for live calls.

Click or use Cmd + Shift + G to toggle.

Glass Cloak always stays above other windows including Zoom, Teams, and Google Meet. It's built into the app — no extra configuration needed.

Hold the mic button or press the shortcut to record. Release to stop and send for transcription. Best for quick, precise dictation.

Hold Space in the floating bar to record.

Click to start recording. Glass Cloak automatically detects silence and stops recording after a configurable pause (default: 2 seconds). Great for natural speech.

Configure silence threshold in Settings → Audio.

Captures system audio continuously — perfect for live meetings. Glass Cloak records all audio output (Zoom, Teams, etc.) and transcribes it in real time.

Whisper runs entirely on your machine — no internet required, no data leaves your device. The Large V3 Turbo model provides excellent accuracy for English and many other languages.

Use OpenAI's Whisper API for faster, more accurate transcription — especially useful on older hardware. Requires an OpenAI API key configured in Settings.

Type Your Prompt

Click the input field in the floating bar and type your question. Press Enter to send. You can also use voice input — it will automatically fill the input field.

Streaming Response

The AI responds in real-time with streaming text. The response panel expands below the floating bar. You can scroll through long responses.

Conversation History

All conversations are saved locally in your SQLite database. Access past chats from the sidebar. Each conversation maintains full context for follow-up questions.

File Attachments

Attach screenshots or files to your prompt for context-aware AI analysis. Drag and drop files onto the floating bar, or use the screenshot feature to capture and attach screen regions.

Press the screenshot shortcut to activate area selection. Click and drag to capture a region. The screenshot is automatically attached to your next prompt.

Cmd + Shift + S to capture

Enable auto-capture to take periodic screenshots at set intervals. Perfect for monitoring a live coding session or tracking changes during a call.

Configure interval in Settings → Screenshots.

After capturing, ask the AI to analyze the screenshot. It can read code, identify errors, describe UI elements, and suggest fixes based on what it sees.

Local AI is the default — everything runs on your machine. No API keys, no internet, no data leaves your device. Download the models from the Model Manager and you're set.

Connect your OpenAI API key for access to GPT-4o and Whisper API. Better accuracy and speed, especially on older hardware. Your key is stored locally and never sent to our servers.

Connect your Anthropic API key for access to Claude models. Known for nuanced, thoughtful responses and strong coding ability. API key is stored locally.