Full-Text Search
AIFull-text search is keyword-based search over document content. It matches query terms against indexed text—exact matches, stemmed forms, or phrase matches—rather than semantic meaning. Traditional search engines use full-text search; vector search uses embedding similarity instead.
Vector vs. Full-Text
| Full-text search | Vector search |
|---|---|
| Matches keywords | Matches meaning |
| "invoice total" → exact or stemmed match | "invoice total" ≈ "amount due" |
| Fast for IDs, proper nouns | Better for paraphrases, concepts |
| Lexical | Semantic |
DocLD uses vector search via Pinecone. Metadata filters can approximate keyword scoping. Hybrid search combines full-text and vector for broader coverage.
Related Concepts
Full-text search complements vector search in hybrid search. RAG and top-k work with retrieved chunks regardless of search type.