Configure via tools.profile in openclaw.json.
Per-agent override: agents.list[].tools.profile.
Live Logs
Real-time gateway log stream
— Waiting for gateway connection —
🧠 Agent WorkspaceDrop agents onto the canvas to get started
Agent Types
🦞
Orchestrator
Commands & deploys sub-agents
🤖
Sub-Agent
Autonomous task runner
🔬
Specialist
Domain expert agent
🌐
API Agent
Manages API integrations
💾
Memory Agent
Persistent knowledge store
⚙️
Tool Agent
Executes tools & scripts
📚
Research Agent
Web & data research
💻
Code Agent
Writes & reviews code
Templates
🔬
Research Pipeline
Gather, analyze & report
💻
Code Review
Review, test & deploy
✍️
Content Creation
Write, edit & publish
🚀
Deploy Pipeline
Build, test & deploy
📊
Data Analysis
Collect, process & visualize
From Gateway
Connect to gateway to see live agents
🧠
Agent Workspace
Drag agent types from the palette onto this canvas
🔍 Inspector
Select an agent node to edit its configuration and objectives.
📋 Agent Logs— All agents
After connecting, tap Overview in the nav bar below to return to the dashboard.
Settings
Configure gateway connection and dashboard preferences
🤖 Local AI Engine
Run agents directly from your browser using your own AI API key. Supports OpenAI (gpt-4o, gpt-4-turbo, etc.) and Anthropic (claude-3-5-sonnet, claude-opus, etc.). Keys are stored only in your browser's localStorage and never sent to any server except the AI provider.
🖥️ Local Backend Launcher
Run the backend directly on this machine for full offline operation.
No Railway account needed. The backend runs at ws://localhost:8080.
Install Node.js (LTS)
Download from nodejs.org and install with default settings.
Verify: node --version
Clone or download the repository git clone https://github.com/barbrickdesign/barbrickdesign.github.io.git
Or download the ZIP from GitHub and extract it.
Install backend dependencies
Open a terminal in the repo folder and run: cd backend && npm install
Start the backend server node server-main.js
Or from the repo root: npm start
The server listens on port 8080.
Connect the dashboard
Click ⚡ Connect to Local Backend above — the dashboard will
switch to ws://localhost:8080.
Tip: Use node backend-auto-start.js
for automatic health-monitoring and restarts.
Full guide: LOCAL_BACKEND_HOW_TO.md
⚡ Quick Connect to Railway Gateway
Paste your Railway service URL and token below, then click Connect.
The token must match the GATEWAY_TOKEN environment variable
set in your Railway service.
Deploy the gateway/ folder to a new Railway service, then enter your URL and token above.
Gateway Connection
Dashboard
Security
Session is stored in localStorage. Sessions expire after 8 hours of inactivity.
About
OpenClaw Enhanced Hub — Advanced control panel for OpenClaw gateway.
Connect to a gateway or configure a Local AI key above, then click any button to dispatch an autonomous agent fix. Results appear below each button.
🩺 Diagnose Dashboard
Full health check — identifies broken tabs, missing data, and disconnected agents
🔄 Fix Data Flow
Audits data flowing through all tabs; identifies mock/placeholder data vs real live data
🤖 Reconnect Agents
Re-reads agent manifest, reconnects all sub-agents, and verifies each is responding
🧹 Clear Stale Sessions
Identifies and closes stale/zombie sessions that may be causing data inconsistencies
🛠️ Self-Heal All Issues
Runs the full autonomous self-healing pipeline — diagnose, patch, reconnect, and verify
Agent Response:
Tips & step-by-step setup guides for this dashboard.
💡 Tips & Setup Guide
Human dev layer tasks · autonomous agent fixes · step-by-step instructions
🔴
Gateway not connected
Connect via Settings → Gateway Connection to enable one-click agent fixes. Local AI key also works.
👨💻 Human Dev Layer — Manual Setup Required
These tasks require a human developer. Autonomous agents cannot handle them because they need credentials,
account access, or physical actions outside the dashboard's reach.
🚂Deploy Backend to RailwayManual▼
The gateway backend must be deployed to Railway so agents can operate 24/7 without your computer running.
In Settings → Local AI Engine: choose Anthropic, model claude-opus-4-5, paste key, click Save.
🔐Set GitHub Repository SecretsManual▼
GitHub Secrets allow Actions workflows and the backend to access API keys securely without committing them to code.
Go to your GitHub repo → Settings → Secrets and variables → Actions.
Click New repository secret.
Add these secrets one by one: AGENTR — your Personal Access Token (PAT) for Agent R OPENAI_API_KEY — if using OpenAI on the backend ANTHROPIC_API_KEY — if using Anthropic on the backend GATEWAY_TOKEN — must match the Railway env variable
To create a PAT: GitHub → Profile → Settings → Developer settings → Personal access tokens → Fine-grained tokens. Grant Contents: Read & Write and Actions: Read & Write.
🖥️Run Backend Locally (Offline Mode)Manual▼
Run the full backend on your own machine for zero-cost development and local-only agent operations.
Create a REST API App to get a Client ID and Secret.
Add to GitHub Secrets: PAYPAL_CLIENT_ID PAYPAL_API (secret key)
Set PAYPAL_CLIENT_ID in Railway environment variables as well.
Test with PayPal Sandbox mode before going live.
⚠️ Revenue-critical: Test thoroughly in sandbox mode before switching to production credentials.
🌐Configure GitHub Pages & Custom DomainManual▼
GitHub Pages serves this dashboard publicly. A custom domain makes it professional and easy to remember.
Go to your repo → Settings → Pages.
Under Source, select Deploy from a branch, branch: main, folder: /.
For a custom domain: buy a domain (e.g. Cloudflare), add a CNAME record pointing to barbrickdesign.github.io.
Enter your domain in the Custom domain box in GitHub Pages settings.
Enable Enforce HTTPS once the domain is verified.
🤖 Agent Fixes — One-Click Automation
These issues can be diagnosed and fixed autonomously. Press a button to dispatch the corresponding agent task.
Requires either a connected gateway or a Local AI key in Settings.
🩺Diagnose Full Dashboard HealthAuto-fixable▼
Runs a comprehensive health check on all tabs — gateway connection, agents, sessions, cron jobs, data flow, and API endpoints. Returns a prioritised report of any issues found.
🔄Audit & Fix Data Flow Through All TabsAuto-fixable▼
Audits every tab to determine whether it is displaying real live data or placeholder/mock data. Lists each tab's data status and provides specific connection instructions for any showing stale data.
🤖Reconnect & Verify All AgentsAuto-fixable▼
Re-reads the agent manifest, reconnects all sub-agents including Agent R, and verifies each is responding with a real status (not demo data). Reports any agents that cannot be reached.
🧹Clear Stale Sessions & Reset StateAuto-fixable▼
Identifies zombie sessions (sessions open but not active for >1 hour), clears them, and resets the session counter. Also purges any orphaned cron tasks or failed queue items.
📱Report Mobile UI IssuesAuto-fixable▼
Analyses the current viewport and identifies any layout overflow, truncated text, unreachable buttons, or other mobile UI problems. Returns specific CSS or layout suggestions.
🛠️Run Full Self-Healing PipelineAuto-fixable▼
Triggers the complete autonomous self-healing sequence: diagnose → reconnect agents → flush stale data → rebuild caches → verify all tabs → report results. This is the "fix everything" button.
🤖 Agent Response
📊 Tab Data Flow Status
Shows whether each tab is receiving live data from the gateway. Connect the gateway to see real status.
Advanced Diagnostics
System checks, error analysis, and step-by-step troubleshooting
System & Environment Checks
🔍
Click "Run All Checks" to diagnose your system
Self-Healing Actions
Automated remediation actions. Each button attempts to detect and fix a specific class of problem.
Troubleshooting Guide
📋
Run diagnostics above to see personalized troubleshooting steps
Initial Setup Guide
Follow these steps to connect OpenClaw Enhanced Hub to your Railway gateway for the first time.
Step 1 — Deploy the Gateway to Railway
Clone or download the gateway/ folder from the repository.
Deploy the gateway/ folder as a new service (drag-and-drop or connect GitHub repo).
In Railway → your service → Variables, add GATEWAY_TOKEN with a strong random secret (32+ characters).
Also add ANTHROPIC_API_KEY or OPENAI_API_KEY for the AI agents.
Wait for the deployment to complete (green checkmark in Railway → Deployments).
Step 2 — Connect the Dashboard
Copy the Railway service domain from Railway → Settings → Networking (e.g. abc123.up.railway.app).
Go to Settings → Quick Connect to Railway.
Enter the WebSocket URL as wss://abc123.up.railway.app (must start with wss://).
Enter the exact GATEWAY_TOKEN value set in Railway (case-sensitive).
Click Connect to Railway.
The status badge in the top-right should turn green (Connected).
Step 3 — Verify Everything Works
Dashboard Overview should show agent and session counts after connecting.
Go to Agents to see your configured agents (from openclaw.json).
Go to Tasks and send a simple test message to an agent.
Go to Logs to confirm the live log stream is working.
Return here and click 🔍 Run All Checks — all should be green.
Step 4 — Common Problems & Fixes
Auth failed / invalid token: Ensure the token in Settings exactly matches GATEWAY_TOKEN in Railway Variables. No leading/trailing spaces.
Connection refused / net::ERR_CONNECTION_REFUSED: The Railway service may not be running. Check Railway → Deployments for errors.
WebSocket handshake fails (HTTPS → WSS mismatch): Use wss:// not ws:// for Railway URLs.
Agents show as "0" after connecting: Check your openclaw.json has at least one agent defined in agents.list.
Keeps disconnecting / reconnecting loop: Check Railway service logs for crash errors. Ensure the service has sufficient memory and no startup errors.
crypto.subtle not available: This page must be served over HTTPS (not plain HTTP). Use https:// or test on localhost.
localStorage blocked: Disable browser extensions that block storage, or open in a normal (non-private) window.
Page loads but lock screen reappears: The 8-hour session TTL has expired, or cookies/storage were cleared. Re-enter your passcode.
Error History
All JavaScript errors and unhandled promise rejections captured this session.
✅
No errors captured this session
WebSocket Frame Inspector
Captures raw WebSocket frames between dashboard and gateway. Useful for debugging authentication and request/response cycles.
Note: disable after debugging to avoid memory growth.
Frame capture disabled. Click "Enable Capture" to start.
🛒 Barbrick Design Shop
Digital products delivered instantly after PayPal checkout
Total Revenue
$0.00
Orders Completed
0
Best Seller
–
Available Products
Loading products…
Order Status
Recent Activity
No activity yet.
Adaptive Intelligence Layer
Context-aware AI suggestions that update as you navigate — real-time guidance and repair prompts
Tracked Events
–
Weighted Keys
–
Saved Panel Order
–
Current Tab
–
🧭 Context-Aware Next Steps
Suggestions update automatically as you navigate. Click any item to go there instantly.
Navigate to any tab to see context-aware suggestions…
⚠️ Detected Issues & Repair Prompts
If functionality appears broken, copy the ready-made repair prompt and paste it into your AI assistant.
No issues detected on current tab.
Predicted Next Actions
Loading…
Top Agents
Loading…
Recently Used Tools
Loading…
Predictive Diagnostics
Real-time anomaly detection on API latency, tokens, memory and agent loops
Avg Latency
–
Token Usage
–
Memory
–
Active Alerts
–
Live Sparklines
API Latency
–
Token Usage
–
Memory (MB)
–
Agent Loop Hz
–
Anomaly Alerts
Monitoring… (samples every 5s)
Revenue Intelligence
Predictive revenue modelling, A/B testing and dynamic pricing
30-Day Revenue Forecast
Loading forecast…
A/B Test Results
No active A/B test
Dynamic Pricing
No pricing data
Marketing Copy Generator
Self-Healing Infrastructure
Auto-restart, patch detection, memory rebuild and exponential-backoff reconnection
Auto-Restarts
0
Patches Applied
0
Memory Rebuilds
0
Reconnects
0
System Status
Initialising…
Self-Healing Log
Quantum-Inspired Task Scheduler
Probabilistic routing, weighted agent selection and emergent load balancing
Dispatch Task
Queue: 0 tasks
Agent Pool
Loading agents…
Dispatch Log
Plugin Manager
Auto-discovered plugins injected into the dashboard
Loaded Plugins
Scanning /plugins/…
Payments
PayPal transaction history and revenue overview
Total Revenue
$0.00
Orders Completed
0
PayPal Mode
Sandbox
Today's Revenue
$0.00
Revenue by Product
No sales recorded yet.
Recent Transactions
No transactions yet.
Product Factory
Create and manage digital products with attached PayPal purchase flows
Create New Product
Products
Loading products…
Content Queue
AI-generated drafts grouped by product — approve or reject before publishing
Live revenue metrics, top performers, and pending high-impact actions
Total Revenue
$0.00
Conversion Rate
0%
Active Products
0
Pending Actions
0
🏆 Top Products
No sales yet.
📉 Needs Attention
All products performing well.
🔥 High-Impact Pending Actions
No high-impact actions pending.
💡 Optimization Proposals
No proposals yet.
♟ Agent R — Core Orchestrator
Supreme authority orchestrator with full system access and autonomous overwatch
♟
AGENTR: Prime — Root Orchestrator
Initialising…
CLEARANCE 999
–
Projects
–
Capabilities
–
Managed Systems
–
Version
Active Capabilities
⏳
Loading…
Orchestration Command
Identity & Authority
♟
Loading manifest…
Managed Systems
⚙️
Loading…
Sub-Agent Systems
🤖
Loading…
📱 Phone FarmBridge: initialising…
🛡️ Admin Control Hub
Monitor, manage, and control all user dashboards across the ecosystem
🛡️
Main Control Hub — Ecosystem Administrator
Scanning for user dashboards…
ADMIN ACTIVE
Registered Hubs
0
Online Now
0
Aggregate Revenue
$0.00
Total Sessions
0
Total Agents
0
Out of Sync
0
User Dashboard Registry
📡
No user dashboards found. Open user-dashboard.html to register it here.
Aggregate Sales — All User Dashboards
No sales data yet. User dashboards will report here.
📢 Broadcast Command
No broadcast history.
⚡ Quick Actions
🔄 Version / Sync Status
Run "Check Version Sync" to see status.
Admin Activity Log
--:--:--ADMINAdmin Hub initialized.
🖥️ Devices
Emulator nodes and phone farm nodes — status, capabilities, and quick actions
Loading emulators…
Loading phones…
📞 VoIP Calls
Initiate outbound calls via VoIP apps running on emulator or phone nodes
New Call
Recent Calls
Time
Number
Node
Status
Duration
Outcome
Transcript
No call logs yet
📲 Social Automation
Trigger app actions (DMs, posts, comments) via APK automation on emulator or phone nodes
New Social Task
Recent Social Tasks
Time
App
Action
Node
Status
Result
No social tasks yet
📋 Task Queue & Logs
Unified view of all VoIP call tasks and social automation tasks
Task ID
Number
Node
Status
Started
Duration
Outcome
Loading…
Task ID
App
Action
Node
Status
Created
Result
Loading…
💬 Messaging Hub
Unified cross-platform messaging — control your device mesh from any channel
🖥️Detecting environment…
📱 Owner Device Registration
Enter your phone number to link this device to the messaging mesh. Agents will send status reports and permission requests to this number.
The hub auto-detects whether you are on a mobile device or a remote PC.
Registered Devices
🔌 Messaging Channels
Connect OpenClaw-supported channels. Messages from all platforms appear in the unified feed below.
Loading channels…
📨 Live Message Feed
Loading messages…
📤 Send Message / Command
Quick Commands
🔔 Notification Configuration
Control which events agents are allowed to message you about. Keep agents useful without being overwhelmed.
Channel Notifications
🌐 Ecosystem Hub
Central command for all dashboards, tools, and systems across Barbrick Design & Consciousness Revolution