So I built UsedBy.

Under the hood, we crawl public sources to extract real-world tool usage and stack combinations. On top of that, we use GPT to add contextual understanding, such as project type, team size, and use case, instead of treating tools as flat categories.

The goal is not ranking tools, but helping people understand which tools are actually used together in practice and why. Recommendations are context-aware rather than popularity-based.

Monetization is intentionally lightweight: ads are shown based on user interest signals like likes and stack interactions, not bidding or sponsored rankings.

Happy to answer questions about the crawlers, data model, or how we handle context and recommendations. Feedback very welcome.