### Install Project Dependencies and Start App

Source: https://github.com/saimnadeemdev/mindwhisperai-cluely-free-alternative/blob/main/README.md

Installs project dependencies using npm and starts the application. Ensure Node.js 18+ and Git are installed, and configure your environment variables before running.

```bash
git clone https://github.com/SaimNadeemdev/MindWhisperAI-Cluely-Free-Alternative.git
cd MindWhisperAI-Cluely-Free-Alternative

npm install

# Set up environment (see Configuration below)
# Start the app
npm start
```

--------------------------------

### Install and Run MindWhisper AI

Source: https://context7.com/saimnadeemdev/mindwhisperai-cluely-free-alternative/llms.txt

Clone the repository, install Node.js dependencies, and start the application. Requires Node.js 18+.

```bash
# 1. Clone the repository
git clone https://github.com/SaimNadeemdev/MindWhisperAI-Cluely-Free-Alternative.git
cd MindWhisperAI-Cluely-Free-Alternative

# 2. Install dependencies (requires Node.js 18+)
npm install

# 3. Start the application
npm start
```

--------------------------------

### Install and Run MindWhisper AI

Source: https://context7.com/saimnadeemdev/mindwhisperai-cluely-free-alternative/llms.txt

Follow these steps to set up and launch the MindWhisper AI desktop application. Ensure you have Node.js installed. Configure your environment variables with an API key or Ollama endpoint.

```bash
git clone https://github.com/saimnadeemdev/mindwhisperai-cluely-free-alternative.git
cd mindwhisperai-cluely-free-alternative
npm install
# Configure .env with API key or Ollama endpoint
npm start
```

--------------------------------

### Configure Local AI with Ollama

Source: https://context7.com/saimnadeemdev/mindwhisperai-cluely-free-alternative/llms.txt

Run MindWhisper AI completely offline using Ollama for enhanced privacy. This requires installing Ollama, pulling a model, and configuring the .env file.

```env
# .env — Local AI mode using Ollama (100% offline, zero data retention)
USE_OLLAMA=true
OLLAMA_MODEL=llama3.2
OLLAMA_URL=http://localhost:11434
```

```bash
# Step 1: Install and start Ollama (https://ollama.com)
ollama serve

# Step 2: Pull the desired model
ollama pull llama3.2         # General interview use
ollama pull codellama        # Coding interviews
ollama pull mistral          # Fast, lightweight responses

# Step 3: Launch MindWhisper AI — it will connect to local Ollama instance
npm start

# Expected output on startup:
# ✓ Connected to Ollama at http://localhost:11434
# ✓ Model: llama3.2 loaded
# ✓ Stealth overlay active — invisible to screen recorders
```

--------------------------------

### Real-Time Transcription and Answer Generation Workflow

Source: https://context7.com/saimnadeemdev/mindwhisperai-cluely-free-alternative/llms.txt

Audio is captured, transcribed with low latency (<300ms), and then sent to an AI model upon user command (Ctrl+Enter) for context-aware responses. This enables features like interview practice with AI feedback.

```bash
# How it works end-to-end:
# 1. App captures system audio / microphone input
# 2. Audio stream → transcription engine (speech-to-text, <300ms)
# 3. Transcript displayed in overlay in real time
# 4. User presses Ctrl+Enter → transcript sent to AI model
# 5. AI response appears in overlay within seconds

# Example — Behavioral interview question:
# Transcript: "Tell me about a time you handled a conflict with a teammate."
#
# AI Response (STAR format):
#   SITUATION: At my previous role, two engineers disagreed on architecture...
#   TASK: I needed to mediate and reach a technical decision before sprint end...
#   ACTION: I scheduled a whiteboard session, listed tradeoffs objectively...
#   RESULT: Team aligned on microservices approach; shipped on time with 0 regressions.
```

--------------------------------

### Configure Supported AI Models

Source: https://context7.com/saimnadeemdev/mindwhisperai-cluely-free-alternative/llms.txt

Select your preferred AI provider by setting environment variables in the `.env` file. Different models are optimized for various tasks, from vision analysis to complex reasoning and local offline use.

```env
# --- Google Gemini (default, cloud) ---
GEMINI_API_KEY=your_key
# Best for: Vision analysis, fast responses, multimodal tasks

# --- OpenAI GPT-5 (cloud) ---
OPENAI_API_KEY=your_key
AI_PROVIDER=openai
AI_MODEL=gpt-5
# Best for: Complex reasoning, multi-step problem solving

# --- Anthropic Claude 4 (cloud) ---
ANTHROPIC_API_KEY=your_key
AI_PROVIDER=anthropic
AI_MODEL=claude-4
# Best for: Nuanced behavioral answers, long-context conversations

# --- Ollama Llama 3.2 (local/offline) ---
USE_OLLAMA=true
OLLAMA_MODEL=llama3.2
OLLAMA_URL=http://localhost:11434
# Best for: Privacy-first usage, no internet required

# --- Ollama CodeLlama (local/offline) ---
USE_OLLAMA=true
OLLAMA_MODEL=codellama
OLLAMA_URL=http://localhost:11434
# Best for: Coding interviews, algorithm problems, debugging

# --- Ollama Mistral 7B (local/offline) ---
USE_OLLAMA=true
OLLAMA_MODEL=mistral
OLLAMA_URL=http://localhost:11434
# Best for: Lightweight, fast responses on low-RAM machines
```

--------------------------------

### Configure Cloud AI (Gemini)

Source: https://github.com/saimnadeemdev/mindwhisperai-cluely-free-alternative/blob/main/README.md

Sets up the Gemini API key for cloud-based AI processing. This is required for using MindWhisper AI's features when not running in local privacy mode.

```env
GEMINI_API_KEY=your_api_key_here
```

--------------------------------

### Configure Cloud AI with Gemini

Source: https://context7.com/saimnadeemdev/mindwhisperai-cluely-free-alternative/llms.txt

Set up MindWhisper AI to use Google Gemini 2.0 Flash for cloud-based responses. Ensure your GEMINI_API_KEY is correctly set in the .env file.

```env
# .env — Cloud AI mode using Google Gemini
GEMINI_API_KEY=AIzaSyD_your_actual_gemini_api_key_here

# Optional: specify model variant
# Default model: gemini-2.0-flash
```

```bash
# After saving .env, start the app — Gemini mode is active by default
npm start

# Expected: App window launches as translucent overlay
# Gemini 2.0 Flash will handle all answer generation and vision analysis
```

--------------------------------

### Trigger Screenshot Analysis with Vision AI

Source: https://context7.com/saimnadeemdev/mindwhisperai-cluely-free-alternative/llms.txt

Use the hotkey Ctrl+H (Win/Linux) or Cmd+H (macOS) to capture screen content. The AI can then interpret code, documents, charts, and error messages for debugging, summarization, or analysis.

```bash
# Trigger screenshot analysis:
# Press Ctrl+H (Win/Linux) or Cmd+H (macOS)

# Supported analysis types:
#   - Code snippets  →  debug, explain, or optimize
#   - Error messages →  root cause analysis + fix suggestion
#   - System design diagrams  →  architectural critique
#   - Documents / PDFs  →  instant summarization
#   - Data tables / charts  →  interpretation and insights

# Example: Screenshot of a Python traceback
# Input (captured screen):
#   Traceback (most recent call last):
#     File "app.py", line 42, in process_data
#       result = data['key']
#   KeyError: 'key'
#
# AI Response generated (Ctrl+Enter):
#   The error is a KeyError on line 42. The dictionary 'data' does not
#   contain the key 'key'. Fix: use data.get('key', default_value)
#   or check 'if "key" in data' before access.
```

--------------------------------

### Configure Local AI (Ollama)

Source: https://github.com/saimnadeemdev/mindwhisperai-cluely-free-alternative/blob/main/README.md

Configures the application to use Ollama for local, privacy-focused AI processing. Specify the model and URL for your Ollama instance.

```env
USE_OLLAMA=true
OLLAMA_MODEL=llama3.2
OLLAMA_URL=http://localhost:11434
```

--------------------------------

### MindWhisper AI Keyboard Shortcuts

Source: https://context7.com/saimnadeemdev/mindwhisperai-cluely-free-alternative/llms.txt

Global hotkeys for controlling MindWhisper AI functions without interrupting workflow. Includes shortcuts for Windows/Linux and macOS.

```text
# Windows / Linux
Ctrl + B          → Toggle overlay visibility (show/hide the window)
Ctrl + H          → Take screenshot for Vision AI analysis
Ctrl + Enter      → Submit current question and get AI response
Ctrl + Arrow Keys → Reposition the overlay window on screen
Ctrl + Q          → Quit the application

# macOS
Cmd + B           → Toggle overlay visibility
Cmd + H           → Take screenshot for Vision AI analysis
Cmd + Enter       → Submit question / get AI response
Cmd + Arrow Keys  → Move overlay window
Cmd + Q           → Quit application
```

```bash
# Example workflow during a technical interview:
# 1. Interviewer asks: "Implement a binary search in Python"
# 2. Press Ctrl+H  →  Screenshot captures the problem on screen
# 3. Vision AI (OCR) reads the problem text automatically
# 4. Press Ctrl+Enter  →  AI generates a complete coded solution
# 5. Press Ctrl+B  →  Hide overlay while you type the answer
# 6. Press Ctrl+B again  →  Show overlay to reference the solution
```

--------------------------------

### Stealth Mode Functionality

Source: https://context7.com/saimnadeemdev/mindwhisperai-cluely-free-alternative/llms.txt

Information on MindWhisper AI's stealth mode, ensuring the always-on-top translucent window is invisible to screen sharing and recording tools. Details platform-specific window flags.

```text
# Stealth mode is enabled by default on launch
# Compatible platforms where the overlay is NOT visible to other participants:
#   - Zoom (Windows, macOS, Linux)
#   - Microsoft Teams
#   - Google Meet
#   - Webex
#   - Any OBS-based recording setup

# The overlay uses platform-specific window flags:
#   Windows: WS_EX_LAYERED + SetLayeredWindowAttributes (transparent to capture)
#   macOS:   NSWindowCollectionBehaviorStationary + CGWindowLevel override
#   Linux:   _NET_WM_STATE_SKIP_TASKBAR + compositor bypass hints
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.