1-based line number where the first match starts. 0 when no match.
Original-case tokens that matched (subset of the query tokens).
Vault-relative path of the matching note (e.g. "Reference/Foo.md").
Total occurrences of all matched tokens. Sort key (desc).
~120-char excerpt centered on the first matched token, with … truncation.
A single hit from searchText.
Hits expose the surrounding snippet and the 1-based line where the first matched token landed so the agent can scroll a UI directly to the relevant passage.
scoreis the total per-token occurrence count (higher = more matches), not normalized — compare scores within the same response only.