In simple English
This program helps you find YouTube videos that got way more views than their subscriber count suggests. You type in a topic, and it tracks down these hidden hits. It then analyzes their titles and scripts to show you what worked. Your API keys and research history stay entirely on your own computer. It costs you no money to run, other than your own YouTube and AI provider usage.
What it is
A local-first web application built with a FastAPI backend and a React frontend. It queries the YouTube Data API to calculate outlier scores and uses a pluggable artificial intelligence model to generate research summaries.
What it is good for
Content creators and researchers who want to reverse-engineer successful YouTube videos without paying for hosted software services.
- Runs entirely on your machine with local data storage.
- Supports thirteen different AI adapters, including local models via Ollama.
- Excludes promoted or artificial traffic by checking engagement ratios.
- Requires Python 3.10 plus and a specific Node version.
- Needs your own YouTube Data API key and AI model key.
- Consumes daily YouTube API quota during searches.
manual YouTube competitor spreadsheets
Ollama
- facebook/docusaurus67k stars
A static site generator from Meta that turns Markdown files into a documentation website.
- mkdocs/mkdocs23k stars
Turns a folder of Markdown files into a documentation website, controlled by one settings file.
- squidfunk/mkdocs-material28k stars
Turns a folder of Markdown files into a searchable documentation website, with every formerly sponsor-only feature now included.
git clone https://github.com/shkuratovdesigner/yuben-app.git cd yuben-app make install make start