Skip to main content

Built-in Subject Types

The Memory Platform provides several built-in subject types that cover the most common entity categories. Using these types ensures compatibility with standard extraction and retrieval behaviors.

Standard Subject Types​

user​

Individual people (customers, employees, stakeholders).

  • ID Pattern: Email, internal UUID, or social ID.
  • Common Memories: Preferences, contact details, interaction history.
  • Graph Node: User

org​

Organizations, companies, teams, or departments.

  • ID Pattern: Domain name or internal slug.
  • Common Memories: Business facts, policies, procurement rules.
  • Graph Node: Org

project​

Specific initiatives, workstreams, or goals.

  • ID Pattern: Project code or UUID.
  • Common Memories: Deadlines, stakeholders, status summaries, decisions.
  • Graph Node: Project

document​

Digital documents, files, or specific pieces of content.

  • ID Pattern: File hash or repository ID.
  • Common Memories: Key findings, executive summaries, author info.
  • Graph Node: Document

tool​

External software, integrations, or agents.

  • ID Pattern: Tool name or API key hash.
  • Common Memories: Performance metrics, usage instructions, capabilities.
  • Graph Node: Tool

channel​

Communication channels (Slack channels, email threads, Discord servers).

  • ID Pattern: Workspace ID + Channel ID.
  • Common Memories: Participation rules, topic summaries.
  • Graph Node: Channel

topic​

Abstract concepts, themes, or categories of knowledge.

  • ID Pattern: Slugified tag (e.g., ai-safety, budget-2025).
  • Common Memories: Definitions, related links, overall insights.
  • Graph Node: Topic

When to Use Built-in Types​

You should use built-in types whenever your entity fits one of the categories above. The platform's Extraction Engine and MCP Tools are optimized to recognize and retrieve these types with higher precision.

If your entity doesn't fit (e.g., a specific "Tenant" or "Legal Case"), see the section on Custom Subject Types.