Getting Started

Core apps to try first

• Chat — Conversational AI with streaming responses. Code blocks render with syntax highlighting. The agent remembers context across your session.
• File Explorer — Browse, read, and edit files with syntax highlighting. The agent can read and write files here too.
• Browser — A built-in web browser with server-side proxy. Ask the agent to browse a URL and it will navigate, read content, and take screenshots.
• Canvas — Visual design surface for building layouts. Drag blocks, arrange slabs, create apps from a prompt.
• Home — Dashboard with activity feed, system status, and quick-launch tiles.

Try the agent's tools

The agent has access to tools that let it interact with your workspace. These are not just chat responses — the agent can take real actions. Try these prompts:

The agent opens the built-in browser, navigates to the URL, reads the page content, and returns a structured summary. You can see the browser window update in real time.

The agent reads the file from the file system and gives you a summary. It sees the same file system you see in the File Explorer.

The window manager

The workspace is a spatial window manager. Windows float, stack, minimize to the taskbar, and restore. You can have multiple windows open at once. The agent can open and arrange windows too.

• Drag window title bars to move them
• Click the taskbar to restore minimized windows
• Start menu groups apps into Workstation and Mission Control sections
• Hash routing updates the URL so you can share workspace state

Keyboard shortcuts

What's next

• Explore all 27 app surfaces — see what each one does
• Understand the agent brain — how tool calls work, what the policy engine does
• Learn about governance — RBAC, entity model, three-ring architecture
• Read the architecture guide — Rust backend, React frontend, DynamoDB