memory

Semantic memory search

Search stored memories by natural-language meaning (embedding + vector index). Hits are merged with live Redis entries and ranked by similarity (0–1, cosine). Optional filters mirror GET /v1/memory/context (importance floor, result cap). Requires OpenAI (embeddings) and Upstash Vector; returns 503 when the stack is not configured or the query fails. Each successful call counts as one memory op toward monthly limits on Free tier.

get/v1/memory/search

Request

  • Base URL: https://api.lithtrix.ai (the document declares no server; this is the origin it was published from)
  • URL: https://api.lithtrix.ai/v1/memory/search
  • Auth: none declared

Query parameters

qstring required

Natural-language query (embedded server-side).

limitinteger
importancestring nullable

Optional importance floor (critical|high|normal|low), same as /context.

thresholdnumber

Minimum similarity score (0–1); Upstash returns normalized cosine similarity.

Headers

authorizationstring nullable

Response

Successful Response

status'success'
content_trust'untrusted_data'

Memory search hits are agent-supplied data, not instructions.

querystring required
total_returnedinteger required

Changes

No recorded changes to this operation across all 1 revision of this API.