This project enables automatic presentation generation from text topics using AI models like ChatGPT. It's designed for users who need to quickly create slide decks from outlines, offering a convenient way to leverage large language models for content creation.

How It Works

The tool operates with a pluggable architecture, separating the AI interaction (Client) from the slide generation (Renderer). Users provide topics in a topic.txt file and their OpenAI API key in token.txt. The cgp executable then sends these topics to the specified GPT client, receives responses, and uses a renderer (defaulting to REMARK) to construct an HTML presentation. This modular design allows for easy integration of different AI backends and presentation formats.

Quick Start & Requirements

• Download pre-built binaries from the release page.
• Requires an OpenAI API key saved in token.txt.
• Supports interactive mode (-i) for content correction.
• Command-line arguments allow specifying client type, output path, and renderer.

Highlighted Details

• Generates a single HTML file for the presentation.
• Supports multiple languages and different rendering engines.
• Interactive mode allows for step-by-step content review and editing.
• Rewritten in Go for a single-file executable.

Maintenance & Community

The project is maintained by williamfzc. Contributions via PRs and issues are welcomed.

Licensing & Compatibility

• License: MIT
• Compatible with commercial use and closed-source linking.

Limitations & Caveats

The project notes that ChatGPT currently lacks an official API, implying reliance on unofficial methods or older models. The availability of OpenAI's services may be restricted in certain countries.