All docs

Workspaces

Organize your knowledge into isolated project contexts with workspaces.

What are workspaces

A workspace is an isolated container for notes, folders, and tags. Each workspace acts like a separate knowledge base — useful when you want to keep different projects, clients, or areas of work cleanly separated.

Every account starts with a default workspace. You can create additional workspaces at any time. If you delete a non-default workspace, its notes and folders are moved back to the default workspace automatically.

Creating and switching workspaces

You can manage workspaces from the sidebar switcher or from Settings:

  • Use the workspace switcher in the sidebar to see all your workspaces and switch between them
  • Create a new workspace from the switcher or settings page — give it a name, optional description, and icon (emoji)
  • A unique URL slug is auto-generated from the workspace name
  • The default workspace cannot be deleted, but it can be renamed

How scoping works

When you select a workspace, everything you see is scoped to that workspace:

  • Notes — only notes belonging to the active workspace appear in the sidebar and search results
  • Folders — the folder tree shows only folders in the current workspace
  • Tags — tag lists and filters reflect only the active workspace
  • Search — full-text, semantic, and hybrid search are all scoped to the active workspace
  • Graph view — the knowledge graph shows connections within the current workspace

Using workspaces via MCP

AI agents can interact with workspaces through MCP tools. This is particularly useful for multi-project workflows where each project has its own knowledge base.

ToolDescription
list_workspacesList all workspaces with note and folder counts
create_workspaceCreate a new workspace, optionally with a template
set_active_workspaceSet the active workspace for the MCP session by ID or slug
get_workspace_contextGet rich context: metadata, folder tree, recent notes, top tags, pinned notes

A typical AI agent workflow looks like this:

  1. Call list_workspaces to discover available workspaces
  2. Call set_active_workspace to switch to the project workspace
  3. Call get_workspace_context to orient and understand what is in the workspace
  4. Subsequent calls to search_notes, create_note, etc. are automatically scoped to the active workspace

See the MCP setup guide for full tool reference and configuration instructions.

Workspace template

When creating a workspace (via MCP's setup_workspace tool), the remedy-pod template is applied by default. It delivers the full Compound Development experience — the same methodology used to build m-notes:

  • 10 AI agent definitions — discovery, spec-writer, planner, designer, marketer, architect, frontend, backend, devops, qa
  • Agent communication protocol — task IDs, 600-token output cap, handoff summaries
  • Story-first workflow — no implementation without an issue branch and acceptance criteria
  • Memory system — context, decisions, stack-notes, and retro templates

The template creates _project/ and _agents/ folders pre-populated with starter notes for each agent and workflow.