Actions22
- Chat Actions
- Container Tag Actions
- Conversation Actions
- Document Actions
- Memory Actions
- Profile Actions
- Search Actions
Search → Search
AI-generatedSummary
Perform a semantic search across stored memories and document chunks using a natural language query, returning the most relevant results.
Inputs
- query (required) — Natural language query to search for. Supermemory returns the most semantically relevant memories and document chunks.
- searchAdditionalFields — Additional options to refine the search (all optional).
Output shape
Array of search result objects, each containing id, content, score, and metadata for matched memories or document chunks.
The operation maps to a POST request to /v4/search with parameters: q (query), containerTag, searchMode, limit, rerank, threshold, and filters derived from filtersJson.
Examples
Example 1: Find recent decisions about pricing
Query: 'What were the key decisions from last week?', Search Mode: Hybrid (default), Limit: 10, Threshold: 0.6
