Roadmap

Current State: The Free Preparation Partner

Ghost currently operates as a powerful, free tool designed to assist with interview preparation. Key features include:

• AI-powered problem extraction from screenshots.

• Solution generation with complexity analysis.

• Debugging assistance.

• A discreet interface for focused practice.

• User choice of AI models (OpenAI/Gemini) using personal API keys.

• Emphasis on privacy with local processing.

This free version will continue to be developed and supported, focusing on helping users learn and practice effectively before actual interviews.

Future Vision: The "Cheater" Tier (Paid Plan)

To explore the cutting edge of real-time AI assistance and support the project's development, we are planning an optional paid tier named "Cheater". This tier will introduce advanced capabilities designed for experimental and research purposes into real-time interaction assistance.

Target Features for the "Cheater" Tier:

1. AI Eye Tracking:

   • Goal: To simulate a consistent gaze towards a specific screen area, regardless of the user's actual eye movement.

   • Mechanism: Utilizes webcam feed and AI analysis to subtly adjust or overlay eye positioning in the video output shared via screen-sharing apps.

   • Purpose: Explores reducing potential suspicion cues related to looking away during high-pressure situations (intended for research/experimental use).

2. Dual Voice Recognition:

   • Goal: To capture and process audio input from both the user (via microphone) and the interviewer (via system audio output).

   • Mechanism: Requires advanced audio routing and real-time speech-to-text processing for both streams simultaneously.

   • Purpose: Enables the AI to understand the full context of the verbal conversation.

3. Real-time Context Awareness & Assistance:

   • Goal: To actively monitor the user's progress and the interview flow (using voice recognition and potentially screen analysis) to provide timely, context-relevant suggestions or answers directly on screen.

   • Mechanism: Combines voice recognition input, potentially screen content analysis, and advanced LLM reasoning to identify moments where the user is stuck or needs information, then generates and displays appropriate text overlays discreetly.

   • Purpose: Explores the boundaries of real-time AI intervention in conversational and problem-solving scenarios.

Planned Pricing for "Cheater" Tier

• Monthly Subscription: $19 / month or 0.1 Sol

Note: Pricing and specific feature implementation details are subject to change based on development complexity and user feedback.