Community Notes provide a scalable, community-based approach to fact-checking, but automating their generation is difficult. The core problem is a trade-off between coverage—how many notes can be produced—and quality, with current methods often sacrificing one for the other.

Existing approaches, such as summarization-based Supernotes, are note-dependent: they require prior notes to work. This dependency limits their ability to operate in situations where no notes exist yet, reducing coverage.

The new method, GitSearch, takes a different route by using gap-informed targeted search. Rather than relying on existing notes, it identifies gaps in coverage and searches for relevant information to fill them, potentially improving both coverage and quality in automated Community Notes generation.