Concepts

Concept docs explain the nouns Alumia uses across the app and API. Start here when you need to understand how the pieces fit together before following a task-specific guide.

Runtime

• Agents — model, prompt, optional fast mode / reasoning effort, memory, tools, and execution lifecycle.
• Peer messaging — how agents coordinate with each other.
• Tools — built-in and connector-backed capabilities, including discovery sets.
• Docs assistant — cited Q&A from any /docs page (⌘I / Ctrl+I).

Work

• Projects and workspaces — ownership, visibility, and shared project access.
• Workspaces — organizing teams and project membership.
• Canvas — live project space for blocks, files, and agent activity.
• Blocks — structured outputs that agents create and update on the canvas.
• Artifacts — explicit multi-file mini-apps (not the default display path).
• Tasks — org task lists with due dates and assignment.
• Schedules — cron-driven agent runs per workspace.
• Shop — store browsing, cart, checkout, and order tracking.
• Feedback — user reports, security probes, workflow triage, and admin handling.

Context

• Memory — persisted facts recalled by agents.
• Knowledge — versioned reference folders and search.
• Datasets — structured tables for agent queries.
• Files — uploads, project folders, and file tools.
• Sessions — chat history and resumable work.
• Skills — platform expansion packs and custom skill bundles.
• Voice — recording, transcription, and realtime agent voice.
• Domains — domain purchase, DNS, and WHOIS profiles.
• Notifications — in-app alerts and preferences.
• Localization — interface language across 25 locales (RTL Arabic) and the separate agent response language.
• MCP servers — org-registered MCP endpoints.
• Learning mode — guided workflows for learning with agents.

Governance

• Admin operations — platform-level governance, connector events, wallet operations, and triage workflows.