This project provides an on-device teleprompter designed to enhance user charisma by suggesting quotes. It targets individuals seeking to improve their public speaking or conversational skills, offering real-time prompts for more engaging delivery.

How It Works

The teleprompter utilizes semantic embeddings to identify relevant quotes based on user input or context. An experimental fine-tuned language model is also available for more advanced quote generation. The core functionality relies on on-device speech recognition for real-time interaction.

Quick Start & Requirements

• Install via conda or pip.
• Run with python main.py (semantic embeddings) or python main.py llm (fine-tuned model).
• Requires transformers, torch.
• Currently MacOS only due to on-device speech recognition dependency (hear library).

Highlighted Details

• On-device processing for privacy and offline use.
• Experimental fine-tuned LLM option.
• Real-time quote suggestion for improved charisma.

Maintenance & Community

This project appears to be a personal project by Daniel Gross. There are no explicit mentions of community channels, roadmap, or ongoing maintenance efforts.

Licensing & Compatibility

The repository does not specify a license. This means it defaults to all rights reserved, and commercial use or integration into closed-source projects is likely not permitted without explicit permission.

Limitations & Caveats

The project is explicitly stated to work only on MacOS. The fine-tuned LLM functionality is described as "highly experimental." The README suggests adding Whisper for broader speech recognition support, indicating current limitations in this area.