Frequently
asked questions.

InvisibleAI is an AI copilot that listens to your calls, meetings and interviews in real time. It detects questions and shows you invisible answers on your screen — nobody else can see them.

Yes. The assistant's floating window doesn't appear in screenshots, screen recordings, or cameras. It was designed specifically for environments where you need discreet help.

For professionals who need to perform at their best: engineers in technical interviews, salespeople on closing calls, executives in English meetings, and students in classes or exams.

Currently available for macOS (Apple Silicon and Intel). The Windows version is in development and coming soon.

The base profile works offline. Premium AI models (GPT-4o, Claude) and audio streaming require internet.

150,000 AI tokens per day, voice transcriptions with Whisper, chat with Llama 4 Scout, and the Interviews profile unlocked. No credit card required.

During the beta, the Pro plan costs $5/mo (regular price: $10/mo). It includes 300,000 daily tokens, unlimited transcriptions, all profiles and premium models.

Users who subscribe during the beta keep the $5/mo price for life, even after the regular price goes up.

No. It's a one-time monthly payment. Cancel whenever you want with no penalty. Access stays until the end of the paid period.

Yes. You can upgrade to Pro at any time. If you downgrade to free, you keep Pro access until your paid period ends.

Your audio is processed locally or through the providers you choose (Groq, OpenAI, Deepgram). The resume you upload never leaves your device — it's processed 100% in the browser. We don't store transcriptions or recordings.

No. InvisibleAI processes audio in real time but doesn't store recordings. Audio is discarded immediately after processing.

Only the text of detected questions is sent to the AI model to generate answers. You can use local models to keep everything on your device.

The free plan includes Llama 4 Scout via Groq. The Pro plan lets you connect your own API keys from OpenAI (GPT-4o), Anthropic (Claude 3.5 Sonnet), and use Groq with additional models.

InvisibleAI captures your microphone and system audio separately. This lets it tell your voice apart from the other person's and improve question detection.

Yes. InvisibleAI is app-agnostic — it works with any app that plays audio: Zoom, Google Meet, Microsoft Teams, Slack, Discord, and more.

With Llama 4 Scout via Groq, answers are generated in under 1 second. With premium models, it can range from 1–3 seconds depending on complexity.