Memory

Items in one scope of a collection, by key in lexicographic order. Narrow with prefix — key your documents so a prefix is meaningful (order:2026-08:...) and this becomes a range scan rather than a filter.

get/v1/memory/collections/{collection}/items

Request

  • Base URL
    https://api.zavu.dev
  • URL
    https://api.zavu.dev/v1/memory/collections/{collection}/items
  • Auth
    HTTP bearer

Path parameters

collectionstring required

Collection name.

Query parameters

scopestring

Scope as a query-string value: contact:<contactId>, conversation:<conversationId>, the literal project, or omitted for project scope.

Only the SHAPE of the id is validated, never its existence — the partition already carries your project, so an id that matches nothing simply addresses an empty scope rather than returning 404.

Example:contact:jd7x2k3m4n5p6q7r8s9t0abc

Which scope to address: contact:<contactId>, conversation:<conversationId>, or omitted for the project scope.

Note the asymmetry with POST /v1/memory and POST /v1/memory/search, which take the scope in the request BODY as scope plus contactId/conversationId. Every other Memory endpoint takes it here.

prefixstring

Only keys starting with this string.

limitinteger
cursorstring

Response

Items in the scope.

nextCursorstring nullable

Example response

{
  "items": [
    {
      "key": "ORD-12345",
      "value": {
        "status": "shipped",
        "total": 42.5
      },
      "rev": 3
    }
  ]
}

Changes