Overview

The OpenClaw WebUI (officially called Control UI or Gateway Dashboard) is the built-in browser interface for OpenClaw. It is a small single-page application built with Vite + Lit and served directly by the OpenClaw Gateway on the same port as the WebSocket server (default: http://<host>:18789/).

No separate installation is required – it appears automatically once the Gateway is running and the UI assets are present. It provides a familiar chat experience plus full control over agents, tools, channels, configuration, and monitoring.

Key Features

Chat Interface: Talk to your OpenClaw agent exactly like ChatGPT (session-based, real-time streaming)
Dashboard Controls: Manage agents, tools, skills, events, and channels
Onboarding Wizard: First-time setup, device pairing, and quick configuration
Quick Open (Local): Instant access from localhost
Device Pairing: Secure first-connection flow
Multi-Language Support: UI available in multiple languages
Live Monitoring: View agent status, memory, cron jobs, and activity
Canvas & Media: Integrated support for A2UI Live Canvas and media skills

Access & URLs

Default URL: http://127.0.0.1:18789/ (or configured basePath)
Remote Access: Strongly recommended via Tailscale Serve (HTTPS, identity-based) or Tailscale Funnel (public with password)

Configuration:

{  "gateway": {    "controlUi": {      "enabled": true,      "basePath": "/openclaw"    },    "bind": "loopback",    "tailscale": { "mode": "serve" }  }}

Security Best Practices

Never expose the WebUI directly to the public internet.
Use Tailscale Serve/Funnel or SSH tunnels only.
Authentication is enforced (token, password, or Tailscale identity).
Set allowedOrigins explicitly for any non-loopback deployment.
Run openclaw doctor regularly to check for misconfigurations.

Use Cases

Local desktop control of your AI assistant
Secure remote access from phone or laptop via Tailscale
Quick testing and debugging of agents/tools without messaging apps
Onboarding new users with the visual wizard
Monitoring long-running autonomous agents

What It Can Do Today

Full chat + configuration, agent lifecycle management, tool testing, event logs, and basic workspace overview. Advanced features (voice, canvas, multi-agent orchestration) are fully controllable through the UI.

For complete documentation and latest updates see the official docs and GitHub repository.

OpenClaw WebUI

Decision information

Security signals (informational, not a security certification)