AI agents · Desktop application
Organise work and supervise AI agents while staying in control.
Alfred combines text and voice conversations, memory, tasks and supervised missions. A 3D scene makes agent activity visible.
Tasks, conversations and project information are spread across tools. Recovering context and identifying priorities means constantly switching applications. The goal was an assistant that organises work and supports execution while keeping the user in control.
A desktop assistant combining text and voice chat, memory for preferences and decisions, tasks, reminders and dependencies. Missions have goals, success criteria and execution limits, with AI spending tracked. A scene created in Blender and integrated with Three.js visualises agent activity. Access to emails and messages requires the user’s authorisation.
Separating proposed, authorised and completed actions. Missions must pause and resume in the background without blocking chat or repeating actions with uncertain outcomes. Memory needs traceable sources, and 3D characters must stay synchronised with server events.
Conversations, organisation and supervised missions in one environment. Users define an objective, approve its launch, track progress and can pause or stop the work. Approval controls, history and the 3D view make actions easier to follow and verify.
Explore the application’s interfaces and user journeys.
