OpenClaw Dashboard: Complete Guide to the Control UI (2026)
Everything about the OpenClaw Dashboard (Control UI) — what it does, how to access it, every feature explained, community alternatives, and remote access setup.
The OpenClaw Dashboard — officially called the "Control UI" — is a browser-based web application that ships with every OpenClaw installation. It runs on the Gateway process, served on port 18789, and gives you a visual interface for managing your entire OpenClaw setup.
If you've been doing everything through config files and the CLI, you're working harder than you need to. The dashboard puts agent management, chat, cron jobs, session logs, and debugging tools in one place.
This guide covers every feature in the dashboard, how to access it locally and remotely, known limitations, and community-built alternatives that fill the gaps.
What the OpenClaw Dashboard Looks Like
The dashboard is built with Vite and Lit, so it loads fast and feels responsive. When you open it, you'll see:
- A header bar reading "OPENCLAW GATEWAY DASHBOARD" with a green "Health OK" indicator (or red if something is wrong)
- A left sidebar with navigation sections
- A main content area that changes based on which section you've selected
Simon Willison described it well: the dashboard is "a whole load of different debugging tools plus a web chat interface." That's accurate. It's not just a status page — it's a full control panel for your agent.
The URL is always http://127.0.0.1:18789/ on a local installation.
Every Feature Explained
The sidebar organizes everything into four main sections: Control, Chat, Agent, and Settings, plus a Resources section. Here's the full breakdown:
| Section | Sub-items |
|---|---|
| Control | Overview, Channels, Instances, Sessions, Cron Jobs |
| Chat | Direct gateway chat session |
| Agent | Skills, Nodes |
| Settings | Config, Debug, Logs |
| Resources | Docs |
Let's go through each one.
Control: Overview
The Overview page is your at-a-glance health check. It shows gateway health status, an agent hierarchy tree displaying parent-child relationships between your main agent and any sub-agents, and live/idle status indicators for each.
If you're running a multi-agent setup, this is where you verify that all agents are connected and responding. The hierarchy tree makes it easy to spot which sub-agent belongs to which parent.
Control: Channels
The Channels view shows every platform connection your agent has — Discord, Telegram, WhatsApp, Slack, or any other configured channel. You can see connection status and channel-specific configuration from here.
Control: Instances
Instances lists the running agent instances tied to your gateway. This is useful when you need to verify that your deployment is actually active and connected.
Control: Sessions
Sessions is one of the most useful pages in the dashboard. It logs every conversation across all channels — Discord, Telegram, WhatsApp, Slack, and the built-in web chat.
For each session, you can see:
- Which model was used
- Context usage as a percentage (how much of the context window was consumed)
- Token count for the session
This is invaluable for tracking costs and debugging. If a user reports a weird response, you can pull up the exact session, see the full conversation, and check whether context limits were the issue.
Control: Cron Jobs
The Cron Jobs page lets you manage scheduled tasks without touching config files. You can:
- List all existing cron jobs
- Add new cron jobs
- Edit existing schedules and prompts
- Run a job immediately (manual trigger)
- Enable/disable jobs without deleting them
- View execution history for each job
There's also a webhook mode that generates a URL with a bearer token, so external services can trigger your cron jobs via HTTP. This is handy for connecting OpenClaw to CI/CD pipelines, monitoring alerts, or other automation tools.
Chat
The Chat section is a full WebChat interface for talking directly to your agent through the gateway. It's not a stripped-down test tool — it supports:
- Streaming responses (tokens appear as they generate)
- Abort/stop button to cancel generation mid-response
- Per-session model overrides (switch models without changing config)
- Search through conversation history
- Pin important messages
- Export conversations
If you're testing prompts or debugging agent behavior, the built-in chat saves you from switching to Discord or Telegram every time you want to test a change.
Agent: Skills
The Skills page shows every skill your agent has loaded, with color-coded status indicators:
- Green = ready and functioning
- Yellow = dependency missing (usually an API key or external service)
You can configure API keys directly from this page for skills that need them. This is much faster than hunting through config files to figure out which environment variable a skill expects.
Agent: Nodes
Nodes shows the internal processing nodes in your agent pipeline. This is more of a debugging/advanced feature — useful when you're building custom agent workflows and need to verify that nodes are connected correctly.
Settings: Config Editor
The Config Editor lets you edit openclaw.json without opening a text editor. It offers two modes:
- A form-based UI for common settings (model, temperature, system prompt, etc.)
- A raw JSON editor for everything else
One smart feature: the editor uses a base-hash guard that prevents concurrent edits. If someone else (or another process) modifies the config while you have it open, the dashboard will warn you instead of silently overwriting their changes.
Settings: Debug
The Debug page exposes internal diagnostic data — request/response inspection, timing information, and error traces. When something goes wrong and the logs aren't enough, this is where you dig deeper.
Settings: Logs
The Logs page provides a live tail of your gateway and agent logs. Features include:
- Real-time log streaming
- Filtering by log level, source, or keyword
- Export logs for sharing or archiving
No more SSHing into your server and running tail -f on log files. The dashboard gives you the same thing in your browser with better filtering.
Resources: Docs
A direct link to OpenClaw documentation from within the dashboard. Simple, but convenient when you need to look something up without leaving your browser tab.
How to Access the OpenClaw Dashboard
There are two ways to open the dashboard:
Option 1: CLI command
openclaw dashboard
This automatically opens your default browser to the dashboard URL.
Option 2: Direct URL
Navigate to:
http://127.0.0.1:18789/
First-Time Access from Non-Localhost
If you're accessing the dashboard from a different device on your network (not localhost), OpenClaw requires device pairing approval for security.
The first time a non-localhost device connects, you'll need to approve it:
openclaw devices approve <requestId>
Localhost connections are auto-approved — no pairing needed when accessing from the same machine.
Remote Access: 3 Methods
Running OpenClaw on a remote server and want to access the dashboard from your laptop? There are three solid approaches.
1. Tailscale Serve (Recommended)
The cleanest option. Tailscale Serve gives you automatic HTTPS and access control without exposing anything to the public internet.
openclaw gateway --tailscale serve
This creates a Tailscale-accessible URL that only devices on your tailnet can reach. No port forwarding, no certificates to manage, no public exposure.
2. SSH Tunnel
The classic approach. Forward the dashboard port through an SSH connection:
ssh -N -L 18789:127.0.0.1:18789 user@your-server
Then open http://127.0.0.1:18789/ in your local browser. Traffic goes through the encrypted SSH tunnel. The dashboard thinks you're accessing from localhost, so no device pairing is needed.
This works well but requires keeping the SSH session open. If the connection drops, you lose access until you reconnect.
3. Identity-Aware Reverse Proxy
For more permanent setups, put the dashboard behind an identity-aware reverse proxy (like Cloudflare Access, Authelia, or Authentik). This gives you proper authentication, audit logging, and the ability to share access with team members.
Security Warning
Never expose the OpenClaw dashboard directly to the public internet without authentication. The dashboard has full control over your agent configuration, API keys, and conversation history. An unauthenticated public dashboard is an open door to your entire setup.
Always use one of the three methods above to ensure access is restricted.
Known Limitations
The dashboard is good, but it's not perfect. Here are the current gaps:
Single-agent connection only. The dashboard can only connect to the default/main agent. If you're running multiple agents on the same gateway, there's no built-in way to switch between them in the UI. This is tracked in GitHub Issue #32495, and a multi-agent switcher PR is currently open.
Complex provider auth requires manual editing. While the Config Editor handles basic settings well, some advanced provider authentication setups (custom OAuth flows, multi-provider fallback chains) still require editing the config file directly.
v2026.3.12 dashboard overhaul. The March 2026 release brought significant UI improvements: split tabs for multitasking, a command palette for quick navigation, and an improved mobile layout. If you're on an older version, updating is worth it just for these changes.
Community Dashboards
The official dashboard covers the essentials, but the community has built alternatives that fill specific gaps. Here are three worth knowing about:
| Project | Key Features | Install |
|---|---|---|
| tugcantopaloglu/openclaw-dashboard | TOTP MFA, cost tracking, activity heatmaps, memory browser | node server.js or Docker |
| mudrii/openclaw-dashboard | Zero-dependency, 12 panels, 6 themes, Go binary | brew install mudrii/tap/openclaw-dashboard |
| grp06/openclaw-studio | Gateway connect, agent management, chat, SSE streaming | npx -y openclaw-studio@latest |
tugcantopaloglu/openclaw-dashboard
This is the feature-heavy option. It adds TOTP-based multi-factor authentication (the official dashboard has none), cost tracking that aggregates API spend across providers, activity heatmaps showing when your agent is busiest, and a memory browser for inspecting what your agent remembers.
Install by cloning the repo and running node server.js, or use the included Docker setup for a containerized deployment.
mudrii/openclaw-dashboard
The polar opposite approach: a single Go binary with zero dependencies. It provides 12 monitoring panels and 6 visual themes out of the box. If you want something lightweight that doesn't need Node.js or Docker, this is it.
Install via Homebrew:
brew install mudrii/tap/openclaw-dashboard
grp06/openclaw-studio
OpenClaw Studio focuses on the development workflow. It connects directly to your gateway, provides agent management tools, a chat interface, and SSE streaming for real-time event monitoring. Think of it as a developer-oriented alternative to the official dashboard.
The fastest way to try it:
npx -y openclaw-studio@latest
No installation needed — npx downloads and runs it in one command.
ClawdHost: Dashboard Access Without the Setup
If you're reading this guide because you're trying to get remote dashboard access working on your VPS, there's an easier path.
ClawdHost includes a managed dashboard out of the box. No Tailscale configuration, no SSH tunnels, no reverse proxy setup. You access your OpenClaw dashboard through your ClawdHost account, and it just works.
The plan is straightforward: $29/month, BYOK (bring your own API key). You get a dedicated Hetzner VPS with OpenClaw pre-configured, one-click deploy, and support for all four major platforms — Discord, Telegram, WhatsApp, and Slack.
Your dashboard is accessible from day one, with proper authentication already in place. No device pairing, no port forwarding, no security configuration on your end.
If you'd rather spend your time building your agent instead of configuring infrastructure, check out ClawdHost.
Wrapping Up
The OpenClaw Dashboard gives you a solid browser-based control panel for your agent. The built-in features cover monitoring, configuration, chat testing, cron job management, and log inspection. For remote access, Tailscale Serve is the smoothest option, with SSH tunnels as a reliable fallback.
The community dashboards are worth exploring if you need features the official UI doesn't have yet — especially cost tracking and MFA. And if you want all of this without the infrastructure work, managed hosting handles the setup for you.
The dashboard keeps getting better with each release. The v2026.3.12 overhaul was a big step forward, and the upcoming multi-agent switcher will close the biggest remaining gap.
Related Articles
OpenClaw Use Cases: 15 Things People Actually Build With It (2026)
Real OpenClaw use cases from real users — from morning briefings and smart home control to autonomous QA agents earning $3,840/mo. What 247K+ GitHub stars actually looks like in practice.
Ollama + OpenClaw: Run Your AI Agent Locally for Free (2026 Setup Guide)
How to run OpenClaw with Ollama for $0 in API costs. Covers setup, recommended models, hardware requirements, common issues, and when local LLMs beat cloud APIs.
Is OpenClaw Free? Real Pricing & Costs Explained (2026)
OpenClaw is free and open-source, but running it isn't free. Here's what it actually costs — software ($0), API usage ($10-500+/mo), hosting ($5-29/mo), and the hidden cost nobody mentions.