TubeScout is a recommendation engine designed to help you escape the YouTube echo chamber.
The standard YouTube algorithm often prioritizes watch time, click-through rates, and established channels, creating a "rich get richer" feedback loop. This makes it difficult to find high-quality content from smaller creators.
Why TubeScout is different: Instead of relying on opaque personalization, TubeScout identifies "hidden gems" by calculating a View-to-Subscriber Ratio. If a video has a high number of views relative to the channel's subscriber count, it indicates that the content is valuable enough to travel outside the channel's existing audience base organically.
This tool puts you back in control of your feed, allowing you to find high-signal content based on your specific interests, not what an algorithm thinks you want to watch.
- Algorithm-Free Discovery: Search based on keywords and recency, not watch history.
- Custom Scoring: Ranks videos using a weighted score of Views, Subscriber Ratio, and Recency.
- Web Interface: A user-friendly Streamlit web app with mobile support.
- Command Line Interface: For quick, scriptable searches.
If you want to run this application locally on your machine:
git clone https://github.com/yourusername/TubeScout.git
cd TubeScoutEnsure you have Python 3.9+ installed. Then run:
pip install -r requirements.txtYou will need to acquire a YouTube v3 API key, which you can do so easily here. A helpful video outlining the process can be found here.
Rename config_template.yaml to config.yaml and enter your API key there.
Note: config.yaml is ignored by git to prevent accidental sharing of your credentials.
For an interactive, visual interface:
streamlit run streamlit_app.pyThis will open the app in your browser. You can enter your API key in the sidebar or let it load from config.yaml.
For terminal-based usage:
python3 yt_search_engine.py 'search term 1' 'search term 2' --search-period 10