Skip to main content
GET
JavaScript
Read one memory layer: its type, description, and how many entries it holds. Not the entries, those you list or get separately.
info returns on every read, includeInfo is a no-op here, and it carries entryCount and createdBy. entryCount is the cheap way to see how full a layer is without paging its entries.

Two kinds of layer

spec.type tells you which, and it changes what the rest of spec means: A skills layer is the reusable knowledge you write and assign to variations. An episodic layer is created and owned by Cadenya for an agent’s episodic memory, so it carries an agentId, an episodicKey, and often an expiresAt. You do not create or edit an episodic layer directly.

Fetch by your own ID

Memory layers are top-level workspace resources, so they take an external ID:
A missing ID is a 404.

List memory layers

Every layer, filterable by type.

List memory entries

The keys inside this layer, without the bodies.

Update a memory layer

Rename it, or set an episodic TTL.

Memory layers

Skills vs episodic, and the cascade.

Authorizations

Authorization
string
header
required

Bearer authentication header of the form Bearer <token>, where <token> is your auth token.

Path Parameters

workspaceId
string
required
Example:

"workspace_01HXKD2E5NQM3T9AYWCF133E3Q"

id
string
required

Memory layer ID. Accepts canonical memlyr_… form or external_id: form.

Example:

"memlyr_01HXKD2E5NQM3T9AYWCFFFBMJH"

Response

OK

MemoryLayer is a named container of memory entries that can be composed into an objective's memory cascade. Layers are workspace-scoped resources. The layer type controls how its entries participate in the agent loop — see MemoryLayerType for details.

See "Memory cascade composition" above for how layers compose at lookup time.

metadata
object
required

Standard metadata for persistent, named resources (e.g., agents, tools, prompts)

spec
object
required
info
object