Core Concepts
Last updated September 21, 2026
Understand the Matruno primitives: workspaces, legal matters, workflow definitions, workflow steps, matter tasks, legal review tasks, documents, agents, integrations, and audit history.
The operating model#
Matruno is an AI-powered Legal Matter Operations platform. A workspace is the tenancy boundary for a law firm, in-house legal team, or legal operations group. It owns users, settings, credentials, matters, legal workflows, templates, and integration configuration.
A matter is the durable legal work record: practice area, matter type, client, participants, confidentiality, jurisdiction, documents and evidence, workflow state, reviews, approvals, generated legal documents, timeline, and audit trail. A workflow definition describes how that matter should move. A workflow step is one node in that definition. A matter task records executable work for a workflow run. A human task records attorney review, approval, feedback, or action required from a person.
Documents, notes, participants, notifications, agent execution logs, and audit events attach back to the matter or workspace so counsel can reconstruct the legal work path without relying on hidden model reasoning.
Matruno is not an autonomous lawyer, a legal-advice chatbot, or a generic CRM.
How the concepts relate#
| Concept | Implemented evidence |
|---|---|
| Workspace | Workspace, WorkspaceUser, workspace settings, member APIs, and workspace-scoped auth dependencies |
| Matter | Matter, Document, Note, AuditLog, MatterTask, participants, matter timelines, and matter detail UI |
| Workflow definition | WorkflowDefinition JSONB config, workflow import/export, validation, model settings, and usage routes |
| Workflow step | WorkflowStep rows and JSON step keys with agent_type, depends_on, branches, params, and terminal markers |
| Matter task | MatterTask records created by workflow launch and advanced by approvals or workflow engine state |
| Human task | HumanTask records for feedback, action, approval, and information tasks used as legal review checkpoints |
| Document template | DocumentTemplate records with pdf_visual and html_dynamic template behavior |
| Agent execution log | AgentExecutionLog, ModeratorAgentLog, agent events, and task log routes |
| Integration | Gmail, Telegram, Nango-backed providers, webhooks, and external tool execution traces where configured |
What Matruno Does Not Expose#
Matruno records execution history, logs, events, state transitions, tool calls, and generated artifacts. Do not describe this as access to every step of model reasoning. The docs should use terms such as agent execution log, audit history, and execution history, not "reasoning trail."
The product also does not expose court e-filing, case-law search, or docket integrations.
Prefer explicit execution modes#
Matters can run in manual or auto execution mode. Treat that as a control surface, not a convenience toggle. Manual mode is useful while validating a new legal workflow or attorney-review path. Auto mode makes sense only when the matter lifecycle, routing rules, and exception handling are already trustworthy. Auto mode still pauses for human legal review wherever the workflow requires it.
Use workflow vocabulary consistently
Use "workflow definition" for the saved JSON configuration and "workflow execution" or "matter workflow" for a live run tied to a legal matter.
Was this helpful?