Search

Semantic and quick search over the knowledge base.

📄️Search documents semantically

Body accepts the semantic-search arguments described below.

📄️Pre-warm the embedding cache for a search query

Generates the text and multimodal embeddings for the query and stores them in the embedding cache, so a following `POST /api/v1/search/` call with the same query reads them from cache instead of waiting on the embedding service. Body: `{ query }` with 5-1000 characters. Validation failures answer `400 { error: 'Invalid query' }`.

📄️Quick-search documents

A GET variant of `POST /api/v1/search/` for typeahead-style lookups: it runs the same semantic search with smaller defaults. `limit` and `maxK` default to 10, and `applicationId` optionally scopes the search.