Control UI & Nodes
Use OpenClaw's Control UI, WebChat, Windows Hub, mobile nodes, Canvas, voice, camera, screen, and location surfaces safely.
OpenClaw is easiest to understand when you separate the Gateway from the surfaces that connect to it. The Control UI, WebChat, desktop apps, and mobile nodes are clients of the Gateway.
Control UI
The Control UI is a browser app served by the Gateway, usually at:
http://127.0.0.1:18789/Open it with:
openclaw dashboardIt can:
- chat through the Gateway
- stream tool calls and live activity
- inspect sessions
- show channel status
- manage some runtime settings
- expose node state
- support talk/realtime sessions where configured
Auth and Pairing
Local loopback browser access is usually simple. Remote browser or device access often requires pairing and scope approval.
openclaw devices list
openclaw devices approve <requestId>
openclaw devices revoke --device <deviceId> --role <role>Treat pairing as operator access. A paired browser or node can see and do more than a casual chat sender.
WebChat
WebChat is the Gateway-backed chat surface. It is useful for:
- local smoke testing
- debugging channel-independent agent behavior
- verifying model credentials
- testing slash commands and usage output
- continuing sessions without a third-party chat app
If WebChat works but a chat channel does not, the problem is probably channel auth, routing, formatting, or delivery.
Windows Hub
Windows Hub is the native Windows companion path. It helps with:
- setup
- tray status
- chat
- node mode
- local MCP mode
- desktop-oriented Gateway workflows
Use it when Windows desktop setup is the primary user path. WSL2 and PowerShell installer flows remain useful for operator and developer setups.
Desktop and Mobile Nodes
Nodes are paired devices that expose local capabilities to the Gateway. Depending on platform and permissions, nodes can provide:
- Canvas rendering
- camera capture
- screen capture or recording
- location
- voice wake or talk mode
- platform-specific actions
Inspect node status:
openclaw nodes status
openclaw nodes listCanvas and A2UI
Canvas is an agent-controlled visual surface. It can render structured UI and give the agent a place to show work beyond text chat.
Use Canvas when:
- the output is visual or interactive
- the agent needs to update a live workspace
- text chat is too narrow for the task
- a node is the natural display target
Keep Canvas actions within the same trust model as other node commands.
Voice and Media Surfaces
Voice and media flows depend on platform, provider, and plugin support. They can include:
- inbound voice notes
- text-to-speech
- realtime talk sessions
- voice wake on supported devices
- camera and screen media
Voice is convenient but easy to over-authorize. Confirm what the node can capture and when the agent is allowed to act.
Debugging Surface Problems
Start by isolating the layer:
openclaw gateway status
openclaw dashboard
openclaw devices list
openclaw nodes status
openclaw logs --followIf the Gateway works but the node does not, inspect pairing, device permissions, local app state, and network reachability. If the node works but a tool fails, inspect tool policy and sandbox settings.
Nodes are execution surfaces
A node is not just a notification client. Once paired, it can expose local device capabilities. Approve devices and scopes as carefully as you would approve shell, browser, or file access.
Continue with practice
You have finished the core ideas of Control UI & Nodes.
If you want to turn the idea into something reusable, continue practicing on AgentWay.