Overview - MemoryLake AI

A Workspace is the top-level organizational boundary in MemoryLake. It acts as a namespace that isolates all data within it — projects, documents, conversations, and facts are all scoped to a workspace.

When to Use Workspaces

Create separate workspaces to isolate data between different tenants, teams, or environments. Everything inside a workspace is invisible to other workspaces. Common patterns:

Multi-tenant apps — one workspace per customer
Environment separation — separate workspaces for development, staging, and production
Team isolation — one workspace per business unit

Relationships

Workspace
├── Projects (knowledge containers)
│   ├── Documents
│   ├── Conversations → Facts (project-scoped)
│   └── Project Facts
├── Actor bindings (who can participate)
│   └── Actor Facts (per-user memory)
└── Agent bindings (which agents are available)

Actors and Agents are created globally, then bound to a workspace to operate within it.
Projects live inside a workspace and hold the actual memory data.
Search operates at the workspace level, querying across all projects within it.

Endpoints

Operation	Method	Path
List Workspaces	`GET`	`/api/v3/workspaces`
Create Workspace	`POST`	`/api/v3/workspaces`
Get Workspace	`GET`	`/api/v3/workspaces/{id}`
Update Workspace	`PATCH`	`/api/v3/workspaces/{id}`
Delete Workspace	`DELETE`	`/api/v3/workspaces/{id}`

​When to Use Workspaces

​Relationships

​Endpoints

When to Use Workspaces

Relationships

Endpoints