AI YouTube Shorts Generator
Turn YouTube videos into viral 9:16 shorts using LLM highlight detection and auto vertical cropping.
Extract viral clips from long-form video with local or API execution.
This tool automates the extraction of viral-ready 9:16 vertical shorts from long-form YouTube videos or local files. It is built for content creators, agencies, and developers wanting an alternative to monthly SaaS clipping subscriptions.
What it is
It analyzes long-form videos using Whisper transcription and GPT or Gemini models to detect highlight moments based on hooks, conflicts, and story peaks. You can run it entirely locally with Python, `faster-whisper`, and OpenCV, or use the default MuAPI backend for faster processing.
Who it's for
It suits technical creators and developers comfortable with the command line who want to avoid per-clip credit caps. It does not suit non-technical users looking for a traditional web browser interface without setup.
Key features
Dual Execution Modes
Switch between a managed API backend and a local setup using faster-whisper and OpenCV.
Virality-Aware Highlights
Ranks clips based on hooks, opinion bombs, and emotional peaks rather than random intervals.
Smart Vertical Crop
Auto-crops 9:16 shorts using MuAPI or local OpenCV face tracking with motion smoothing.
JSON Data Export
Dump full transcripts, candidate highlights, and clip metadata via the `--output-json` flag.
Up and running in 3 steps
Our honest verdict
What's great
- No monthly subscriptions or per-clip credit limits on the open-source codebase.
- Fully editable highlight ranking framework instead of a black-box algorithm.
- Supports local file paths and `file://` URLs to bypass YouTube downloading entirely.
- Outputs viral scores, opening hook lines, and explanations for every generated clip.
Where it falls short
- Requires managing Python virtual environments, dependencies, and command-line flags.
- Local mode demands local compute resources for faster-whisper and OpenCV face tracking.
- API mode relies on external MuAPI keys and services.
Pricing & license
The code is free under its license, but running it requires your own API keys for LLMs and Whisper, or a MuAPI key if using API mode.
Open remote AI-training roles
873 open remote AI-training roles, updated November 15, 2026; most pay $20–$170 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
These reviewed AI tools do the same job instantly — from our scored directory.
Use it, or have us set it up
FAQ
Can I run this completely offline?
Yes, using `--mode local`, which uses local faster-whisper, OpenCV, and your own OpenAI or Gemini API key strictly for the LLM highlight step.
What aspect ratios are supported?
The default is 9:16 for TikTok, Reels, and Shorts, but you can specify other aspect ratios using the CLI flags.
