A workspace, not an isolated conversation
Long-running conversations connect with workspaces, rooms and documents. An agent can work with the materials relevant to a task while keeping different areas of life organized.
Structured memory and contextual retrieval
Aru can organize durable user-approved information and retrieve relevant context across conversations. The goal is continuity and useful context without placing every historical record into every request.
Models, tools and system capabilities
Agents can use the models and MCP tools you enable, and—with explicit system permission—work with capabilities such as Calendar, Reminders and HealthKit. This turns a plan into actions within a clear permission boundary.
User-owned context and explicit control
You choose the provider, connected capabilities and information available to Aru. Local storage, scoped retrieval and per-feature permissions keep the agent’s working context understandable and controllable.