Supermemory icon

Supermemory

Interact with the Supermemory API

Actions23

Memory → Search

AI-generated

Summary

Search memory entries (and optionally documents, summaries, related memories, forgotten memories) using a query string.

Inputs

  • query (required) — The search query text
  • containerTag — Optional container tag to scope the search
  • filters — Metadata filters as JSON (supports AND/OR logic)
  • includeDocuments — Whether to include associated documents in results
  • includeForgottenMemories — Whether to include forgotten memories in results
  • includeRelatedMemories — Whether to include related memories for context
  • includeSummaries — Whether to include document summaries in results
  • limit — Max number of results to return (default 50, min 1, max 100)
  • rerank — Whether to rerank results by relevance
  • rewriteQuery — Whether to rewrite query for better matching (adds ~400ms latency)
  • searchMode — Search scope — memories only, documents only, or hybrid (default memories)
  • threshold — Sensitivity for memory selection (0 = least sensitive, 1 = most sensitive, default 0.6)

Output shape

array of objects

Each object represents a search result; when include flags are true, related data such as documents, summaries, related memories, or forgotten memories may be present.

Examples

Example 1: Find recent memories about project X

Set query to 'project X', limit to 10, threshold to 0.7, and enable rerank for better relevance.

Links

Discussion