> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.revvue.ai/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.revvue.ai/_mcp/server.

# WikiSearchHit

GraphQL object

One hit from a hybrid (FTS + vector + trigram) wiki search.

Reference: https://docs.revvue.ai/api-reference/knowledge-hub/types/objects/wiki-search-hit

## Fields

- `page` — Matching page (summary, no body).
- `score` — Combined relevance score; higher = better. Not normalised across queries.
- `matchKind` — Which retrieval track ranked this hit: HYBRID, FTS, SEMANTIC, or TRIGRAM.
- `ftsScore` — Raw Postgres FTS rank score (ts_rank_cd) when the FTS track ranked this hit; None otherwise.
- `semanticScore` — Raw pgvector cosine similarity in 0..1 (1 = identical) when the vector track ranked this hit; None otherwise. Use this as a confidence band: \<0.3 is weakly related, >0.7 is strong.
- `sectionHeading` — Markdown heading of the section most likely to answer the query, without `#` markers. None when the page has no headings or no clear best match.
- `sectionBody` — Body of the matched section (capped at `section_chars`, default 1200; trailing `…` marks truncation). Lets the caller skip a follow-up `wiki_read_page` round-trip for most queries.

## Definition

```graphql
type WikiSearchHit {
  page: WikiPageSummary
  score: Float
  matchKind: SearchMatchKind
  ftsScore: Float
  semanticScore: Float
  sectionHeading: String
  sectionBody: String
}
```