New Term Radar reads where new AI names appear first — X accounts, Hugging Face trending, GitHub breakout repos, competitor sitemaps and Search Console — then maps the generic searches behind them and sizes them against Google Trends. The goal: help people who build pages, posts and products act before keyword tools show the term at all.

The radar engine started as an open-source daily report (github.com/zhuyansen/new-term-radar). This site is the commercial layer on top: the same signal pipeline, with an opportunity scoring system, a public trend board, and research built from real runs.

We publish our known limitations on the methodology page on purpose. Every number shows its source and capture date; every lead gets a 14-day review so we can measure how early we actually were.