> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.cloudraker.com/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.cloudraker.com/_mcp/server.

# Space types

Space types are **admin-only**. If you don't have an admin role, you won't reach this area — you'll be told you need an admin role to manage space types. You can still create spaces from existing types; see [Create a space](/spaces/create-a-space).

A **space type** is a reusable definition that new spaces are created from. Unlike a template that gets copied, a space type stays attached to every space made from it and governs who has access, how the knowledge graph is extracted, and which playbooks can run.

Open space types from **Space types** in the sidebar. Create one with **New space type**, giving it:

* A **Title** — for example, "Finance close".
* An **Identifier** — lowercase letters, digits, and hyphens, starting with a letter and not ending in one, 2–40 characters (for example, `finance-close`).

A space type's detail page has three tabs.

## Settings — members and roles

The **Settings** tab controls the members and roles that apply to **every** space of this type — "Roles granted here apply to every space of this type." Add people with **Add member**; new members start as viewers.

The role ladder mirrors a single space's ([Spaces overview](/spaces/overview)) but adds an extra **Lister** rung at the bottom:

| Role            | What it grants across all spaces of this type |
| --------------- | --------------------------------------------- |
| **No access**   | No inherited access.                          |
| **Lister**      | See that the spaces exist.                    |
| **Viewer**      | Read the spaces.                              |
| **Contributor** | Read, upload, and run.                        |
| **Manager**     | Plus share and manage access.                 |
| **Archiver**    | Plus archive.                                 |

Roles are cumulative. Access granted here is **inherited** by every space of this type and is separate from the direct roles set inside an individual space.

## Ontology — how the knowledge graph is built

The **Ontology** tab ("Ontology extraction") guides how the [knowledge graph](/spaces/search-and-graph) is extracted from files uploaded to spaces of this type:

* **Extraction guidance** — freeform instructions added to the extraction prompt (for example, "Focus on procurement documents; treat SOWs as contracts").
* **Entity types** — an optional custom list. When set, it **replaces** the built-in entity types, and extracted entities are validated against the keys you define. Each entry has a lowercase key and an optional description to guide the model.
* **Provide the current date and time to the model** — a toggle that helps the model resolve relative dates.

Save with **Save ontology settings**. These settings shape what shows up on the **Graph** tab of every space of this type.

## Playbooks — what can run

The **Playbooks** tab lets you "Choose which playbooks can run in spaces of this type." Each row shows the playbook's **Status** (Draft, Active, or Archived) and its **Availability**:

* **Available in all space types** — the playbook is global and runs in every space type.
* **Available in this space type** — the playbook is bound specifically to this type.

Only playbooks made available here can be run in this type's spaces. Playbooks themselves are defined separately; see [Playbooks overview](/playbooks/overview).