Skip to content
By Hidini Team 3 min read

Getting Started with Hidini

guide setup getting-started

What is Hidini?

Hidini is a macOS overlay application that gives you real-time AI assistance during technical interviews. The overlay floats on top of your screen but is completely invisible to screen sharing. Zoom, Google Meet, and Microsoft Teams cannot see it.

You get real-time transcription of the conversation, AI-powered hints on demand, and the ability to screenshot coding problems for instant analysis. All from a floating window that only you can see.

System Requirements

  • macOS 14 Sonoma or later
  • Apple Silicon or Intel Mac
  • Microphone access (for transcription)
  • Screen recording permission (for screenshots)

Installation

  1. Download Hidini from our website
  2. Drag the app to your Applications folder
  3. Launch the app. A small Hidini “H” icon appears in your menu bar
  4. Grant microphone and screen recording permissions when prompted

That’s it. No subscription, no sign-up, and no license key for the free tier.

Choosing an AI Provider

Hidini ships with AI already configured, so you can run your first session without setting anything up. The free tier includes 50 AI credits, and credits are deducted per request, with cheaper models costing fewer credits than the frontier ones.

When you want to stop using credits, you have two options.

Bring your own API key

Connect the provider you already pay for and Hidini calls it directly. No credits are consumed.

  1. Open Hidini settings (click the menu bar icon)
  2. Go to the AI Provider tab
  3. Select your provider: Claude (Anthropic), GPT (OpenAI), or Gemini (Google)
  4. Enter your API key. It is stored in your macOS Keychain
  5. Choose your preferred model

Run a model locally for free

Local models cost nothing and never send your conversation off your machine.

  1. Install Ollama and run ollama pull llama3.1 in your terminal
  2. In Hidini settings, select Local AI as your provider
  3. Leave the endpoint at http://localhost:11434/v1, or point it at LM Studio, Jan, or any other OpenAI-compatible server
  4. Type the model name you pulled

Choosing a Transcription Engine

Hidini captures both your microphone audio and the interviewer’s audio from the video call. You need a transcription engine to convert that audio to text.

EngineCostAccuracyLatency
Apple SpeechFreeBasicLow
Groq Whisper$0.02/hrHigh~80ms
OpenAI Realtime$0.003/minVery HighLow
Deepgram Nova-3$0.46/hrVery HighReal-time

For your first session, Apple Speech works fine. Upgrade to Groq or Deepgram when you want better accuracy.

Your First Interview Session

  1. Launch Hidini and the overlay appears on screen
  2. Join your video call on Zoom, Meet, or Teams
  3. Press Cmd+Shift+L to start transcribing
  4. Watch the transcript build in real time as the interviewer speaks
  5. Press Cmd+Shift+Enter to get an AI hint based on the conversation
  6. Press Cmd+Shift+A to draw a box around a coding problem and send it for AI analysis

The overlay is invisible to the interviewer the entire time. They see your desktop, your code editor, your browser, but never Hidini.

Next Steps

← All posts