memories

Creates, updates, deletes, gets or lists a memories resource.

Overview

Name	memories
Type	Resource
Id	databricks_workspace.agentkit.memories

Fields

The following fields are returned by SELECT queries:

get

Name	Datatype	Description
name	string	Resource name in the form ``memory-stores/{managed_memory_store_id}/entries/{managed_memory_entry_id}``. Required for updates; assigned by the service on creation.
actor_id	string	Customer-provided identifier for the actor whose memory this entry represents.
session_id	string	Optional identifier for the session associated with this memory entry. When omitted, the entry applies across the actor's sessions.
content	string	Optional free-form memory content.
create_time	string (date-time)	Time when the entry was created.
description	string	Human-readable description of the memory entry.
path	string	Absolute, case-sensitive path identifying the entry within its actor and optional session. Paths must begin with ``/`` and must not contain empty, ``.`` or ``..`` segments.
source_type	string	Which writer created this entry. Caller sets this on Create; immutable after creation. (MANAGED_MEMORY_ENTRY_SOURCE_TYPE_AGENT, MANAGED_MEMORY_ENTRY_SOURCE_TYPE_DREAMER)
update_time	string (date-time)	Time when the entry was last updated.

list

Name	Datatype	Description

Methods

The following methods are available for this resource:

Name	Accessible by	Required Params	Optional Params	Description
get	select	managed_memory_store_id, managed_memory_entry_id, deployment_name, workspace_domain	view	Retrieves a memory entry by resource name. view controls whether the response includes memory
list	select	managed_memory_store_id, actor_id, deployment_name, workspace_domain	page_size, page_token, path_prefix, session_id, view	Lists memory entries for one actor. An exact path filters entries across sessions, ignoring
create	insert	managed_memory_store_id, deployment_name, workspace_domain, actor_id, path	managed_memory_entry_id	Creates a memory entry using exclusive-create semantics. Callers may choose the entry ID; the service
update	update	managed_memory_store_id, managed_memory_entry_id, update_mask, deployment_name, workspace_domain, actor_id, path		Updates selected mutable fields on a memory entry. Identity fields are immutable.
delete	delete	managed_memory_store_id, managed_memory_entry_id, deployment_name, workspace_domain		Deletes a memory entry by resource name. Returns NOT_FOUND when the entry does not exist in the
search	exec	managed_memory_store_id, deployment_name, workspace_domain, actor_id, query		Searches memory entries by text query for one actor. Returns matching entries and scores ranked by

Parameters

Parameters can be passed in the WHERE clause of a query. Check the Methods section to see which parameters are required or optional for each operation.

Name	Datatype	Description
actor_id	string	Customer-provided identifier for the actor whose entries are listed.
deployment_name	string	The Databricks workspace deployment name: the host prefix of the workspace URL (dbc-abcd0123-a1bc on AWS, adb-1234567890123456.7 on Azure, 1234567890123456.7 on GCP). Set DATABRICKS_DEPLOYMENT_NAME to default it for the session. (default: dbc-abcd0123-a1bc, x-stackQL-envVar: DATABRICKS_DEPLOYMENT_NAME)
managed_memory_entry_id	string	The managed_memory_entry_id segment of the name resource name (memory-stores/{managed_memory_store_id}/entries/{managed_memory_entry_id}). Resource name in the form memory-stores/{managed_memory_store_id}/entries/{managed_memory_entry_id}.
managed_memory_store_id	string	The managed_memory_store_id segment of the parent resource name (memory-stores/{managed_memory_store_id}). Memory store whose entries are searched, in the form memory-stores/{managed_memory_store_id}.
update_mask	object	Fields to update. Only content and description may be updated.
workspace_domain	string	The Databricks workspace domain suffix for the cloud. Selected by DATABRICKS_CLOUD (aws, azure or gcp; default aws). (default: cloud.databricks.com, x-stackQL-envVarSelector: {"envVar":"DATABRICKS_CLOUD","default":"aws","values":{"aws":"cloud.databricks.com","azure":"azuredatabricks.net","gcp":"gcp.databricks.com"}})
managed_memory_entry_id	string	Optional caller-selected memory entry ID. The service generates an ID when omitted.
page_size	integer	Maximum number of entries to return. The service may return fewer entries than requested. Defaults to 10; must be between 1 and 100.
page_token	string	Opaque pagination token from a previous ListManagedMemoryEntries response.
path_prefix	string	Optional path prefix used to restrict entries within the actor partition.
session_id	string	Optional session identifier. When set, only entries with this exact session_id are returned. Omitted-session (cross-session) entries are not included. Ignored when path is set.
view	string	Controls whether each returned entry includes memory content. Defaults to FULL.

SELECT examples

get

Retrieves a memory entry by resource name. view controls whether the response includes memory

SELECT
name,
actor_id,
session_id,
content,
create_time,
description,
path,
source_type,
update_time
FROM databricks_workspace.agentkit.memories
WHERE managed_memory_store_id = '{{ managed_memory_store_id }}' -- required
AND managed_memory_entry_id = '{{ managed_memory_entry_id }}' -- required
AND deployment_name = '{{ deployment_name }}' -- required
AND workspace_domain = '{{ workspace_domain }}' -- required
AND view = '{{ view }}'
;

list

Lists memory entries for one actor. An exact path filters entries across sessions, ignoring

SELECT
*
FROM databricks_workspace.agentkit.memories
WHERE managed_memory_store_id = '{{ managed_memory_store_id }}' -- required
AND actor_id = '{{ actor_id }}' -- required
AND deployment_name = '{{ deployment_name }}' -- required
AND workspace_domain = '{{ workspace_domain }}' -- required
AND page_size = '{{ page_size }}'
AND page_token = '{{ page_token }}'
AND path_prefix = '{{ path_prefix }}'
AND session_id = '{{ session_id }}'
AND view = '{{ view }}'
;

INSERT examples

create

Creates a memory entry using exclusive-create semantics. Callers may choose the entry ID; the service

INSERT INTO databricks_workspace.agentkit.memories (
actor_id,
path,
content,
create_time,
description,
name,
session_id,
source_type,
update_time,
managed_memory_store_id,
deployment_name,
workspace_domain,
managed_memory_entry_id
)
SELECT 
'{{ actor_id }}' /* required */,
'{{ path }}' /* required */,
'{{ content }}',
'{{ create_time }}',
'{{ description }}',
'{{ name }}',
'{{ session_id }}',
'{{ source_type }}',
'{{ update_time }}',
'{{ managed_memory_store_id }}',
'{{ deployment_name }}',
'{{ workspace_domain }}',
'{{ managed_memory_entry_id }}'
RETURNING
name,
actor_id,
session_id,
content,
create_time,
description,
path,
source_type,
update_time
;

Manifest

# Description fields are for documentation purposes
- name: memories
props:
  - name: managed_memory_store_id
    value: "{{ managed_memory_store_id }}"
    description: Required parameter for the memories resource.
  - name: deployment_name
    value: "{{ deployment_name }}"
    description: Required parameter for the memories resource.
  - name: workspace_domain
    value: "{{ workspace_domain }}"
    description: Required parameter for the memories resource.
  - name: actor_id
    value: "{{ actor_id }}"
    description: |
      Customer-provided identifier for the actor whose memory this entry represents.
  - name: path
    value: "{{ path }}"
    description: |
      Absolute, case-sensitive path identifying the entry within its actor and optional session. Paths must begin with ``/`` and must not contain empty, ``.`` or ``..`` segments.
  - name: content
    value: "{{ content }}"
    description: |
      Optional free-form memory content.
  - name: create_time
    value: "{{ create_time }}"
    description: |
      Time when the entry was created.
  - name: description
    value: "{{ description }}"
    description: |
      Human-readable description of the memory entry.
  - name: name
    value: "{{ name }}"
    description: |
      Resource name in the form ``memory-stores/{managed_memory_store_id}/entries/{managed_memory_entry_id}``. Required for updates; assigned by the service on creation.
  - name: session_id
    value: "{{ session_id }}"
    description: |
      Optional identifier for the session associated with this memory entry. When omitted, the entry applies across the actor's sessions.
  - name: source_type
    value: "{{ source_type }}"
    description: |
      Which writer created this entry. Caller sets this on Create; immutable after creation.
  - name: update_time
    value: "{{ update_time }}"
    description: |
      Time when the entry was last updated.
  - name: managed_memory_entry_id
    value: "{{ managed_memory_entry_id }}"
    description: Optional caller-selected memory entry ID. The service generates an ID when omitted.
    description: Optional caller-selected memory entry ID. The service generates an ID when omitted.

UPDATE examples

update

Updates selected mutable fields on a memory entry. Identity fields are immutable.

UPDATE databricks_workspace.agentkit.memories
SET 
actor_id = '{{ actor_id }}',
path = '{{ path }}',
content = '{{ content }}',
create_time = '{{ create_time }}',
description = '{{ description }}',
name = '{{ name }}',
session_id = '{{ session_id }}',
source_type = '{{ source_type }}',
update_time = '{{ update_time }}'
WHERE 
managed_memory_store_id = '{{ managed_memory_store_id }}' --required
AND managed_memory_entry_id = '{{ managed_memory_entry_id }}' --required
AND update_mask = '{{ update_mask }}' --required
AND deployment_name = '{{ deployment_name }}' --required
AND workspace_domain = '{{ workspace_domain }}' --required
AND actor_id = '{{ actor_id }}' --required
AND path = '{{ path }}' --required
RETURNING
name,
actor_id,
session_id,
content,
create_time,
description,
path,
source_type,
update_time;

DELETE examples

delete

Deletes a memory entry by resource name. Returns NOT_FOUND when the entry does not exist in the

DELETE FROM databricks_workspace.agentkit.memories
WHERE managed_memory_store_id = '{{ managed_memory_store_id }}' --required
AND managed_memory_entry_id = '{{ managed_memory_entry_id }}' --required
AND deployment_name = '{{ deployment_name }}' --required
AND workspace_domain = '{{ workspace_domain }}' --required
;

Lifecycle Methods

search

Searches memory entries by text query for one actor. Returns matching entries and scores ranked by

EXEC databricks_workspace.agentkit.memories.search 
@managed_memory_store_id='{{ managed_memory_store_id }}' --required, 
@deployment_name='{{ deployment_name }}' --required, 
@workspace_domain='{{ workspace_domain }}' --required 
@@json=
'{
"actor_id": "{{ actor_id }}", 
"query": "{{ query }}", 
"limit": {{ limit }}, 
"page_size": {{ page_size }}, 
"page_token": "{{ page_token }}", 
"path_prefix": "{{ path_prefix }}", 
"session_id": "{{ session_id }}", 
"view": "{{ view }}"
}'
;