KV Memory Search API

About this tool

Search all your stored key-value pairs by semantic similarity. When you store a value via /memory/kv/set, an embedding is automatically generated. This endpoint embeds your query and ranks all stored values by cosine similarity — returning the most relevant entries. Perfect for agent long-term memory, knowledge retrieval, and fuzzy lookups.

Quick Start

curl -X POST https://api.iteratools.com/memory/kv/search \
  -H "Authorization: Bearer YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "user interface preferences",
    "namespace": "myapp",
    "limit": 5
  }'

Response

{"ok": true, "data": {"results": [{"key": "user:preferences", "value": {"theme": "dark", "language": "en"}, "score": 0.89}], "count": 1}}

Parameters

Field	Type	Description
`query`	string *	Natural language search query
`namespace`	string	Namespace to search in (default: "default")
`limit`	integer	Max results to return (default: 10, max: 100)

Pricing

$0.002 per search via x402 micropayment on Base (USDC).

Full Documentation Browse All Tools