Overview
MemoryLake organizes knowledge in a two-level hierarchy: workspaces contain projects. Workspaces are the top-level container (your personal space has a default one; teams create more); projects are focused bodies of knowledge inside them — the unit you point an AI at.What is a Project?
A project is:- A retrieval scope — searches, Playground context, and agent recall operate within a project, so contexts don’t bleed into each other
- A container for documents, extracted memories, and Open Data subscriptions
- An integration unit — agent plugins and MCP keys connect to one project
- A permission boundary — in teams, access can be granted per project
A default workspace (with a first project) is provisioned automatically when you sign up — you only manage this hierarchy explicitly when your usage grows.
The Project Dashboard
Open a project (console → Workspaces → workspace → project) to find its tabs:
See Project Dashboard for details.
Core Guides
Project Dashboard
Navigate tabs, monitor processing, manage the project.
Document Association
Link documents from your library into projects.
How to Slice Projects
- Personal: one per life/work context — “Health”, “Home purchase”, “Side project”. Ask: should an assistant answering about X ever cite Y? If no, separate projects.
- Team: one per initiative, client, or domain — “Q3 launch”, “Customer X”, “Compliance 2026”. Project-level permissions let you grant a contractor exactly one project.
- Builders: commonly one project per end user (or tenant) — see Build Apps with Memory.
Common Use Cases
Research Organization
Research Organization
Group papers, notes, and data files in one project; subscribe research Open Data; connect your agent. Questions get answered from exactly this corpus.
Client Engagements
Client Engagements
One project per client. The OpenClaw plugin’s per-directory config pins each client’s working folder to its project — memories never cross clients.
Team Knowledge Domains
Team Knowledge Domains
Projects per domain (“Pricing”, “Legal”, “Onboarding”) inside a shared workspace, with per-project permissions where needed.
Getting Started
1
Create a project
In your workspace, create a project with a clear name and purpose description.
2
Add content
Upload files in the Files tab, associate existing Library documents, subscribe Open Data categories.
3
Chat and accumulate memory
Use the Playground scoped to this project — memories extract automatically. Review them in the Memories tab.
4
Connect integrations
From the Integrations tab, install the OpenClaw plugin or set up API access for this project.