This open-source project addresses the challenge of efficiently repurposing long-form YouTube videos into engaging short-form content. It offers creators a free, automated solution for generating viral clips, complete with highlight detection, subtitles, translation, and voiceovers, serving as a direct alternative to paid SaaS platforms like OpusClip and Vidyo.ai.

How It Works

The system downloads YouTube videos or accepts local files, then utilizes faster-whisper for local, timestamped transcription. A selected Large Language Model (LLM) classifies the video content type to tailor highlight detection. This LLM then analyzes the transcript through a "virality framework"—identifying potential hooks, emotional peaks, and quotable moments—to rank and select the top N highlights. Finally, each highlight is automatically cropped into a vertical format (defaulting to 9:16), with an optional AI-generated hook prepended to the clip.

Quick Start & Requirements

• Primary install: Clone the repository, create a Python virtual environment, and install dependencies via pip install -r requirements.txt.
• Prerequisites: Python 3.10+, an API key for a supported LLM provider (OpenAI, Gemini, or MuAPI). Local Whisper model selection (tiny, base, small, medium, large-v3) impacts performance.
• Configuration: Set the LLM provider and API key in a .env file.
• Usage: Run via CLI: python main.py <youtube_url_or_local_path>. A local web interface is also available (python3 server.py and cd web; python3 -m http.server 8000).
• Links: Gemini: https://aistudio.google.com, OpenAI: https://platform.openai.com, MuAPI: https://muapi.ai.

Highlighted Details

• Free and open-source alternative to commercial short-video generation tools.
• Generates clips without pre-clip credits or watermarks.
• Provides an API for integration into custom projects.
• Features context-aware AI-generated hooks for enhanced engagement.
• Supports multiple LLM providers (OpenAI, Gemini, MuAPI) for flexibility.
• Employs faster-whisper for efficient, local transcription.

Maintenance & Community

The provided README does not detail specific contributors, community channels (like Discord or Slack), or a public roadmap.

Licensing & Compatibility

This project is licensed under the MIT License, which permits broad use, including commercial applications and integration into closed-source projects.

Limitations & Caveats

The README does not explicitly state limitations. However, functionality relies on external LLM API access, which may incur costs and is subject to provider rate limits. The quality of generated highlights and transcriptions depends on the chosen LLM and Whisper model configurations.