Rem Dashboard

A native Tauri dashboard for monitoring OpenClaw AI assistants. Designed for small displays (5-inch / 1280×720) on always-on setups like Mac Mini.

Features

• 🌸 Token Usage — Petal visualization for 5-hour and weekly Claude Max limits
• 🤖 Agent Status — Real-time primary agent activity with auto-detection
• ⏰ Cron Jobs — View launchd scheduled tasks with next run times
• 📋 Activity Feed — Color-coded logs via Server-Sent Events (SSE)
• 🫧 Sub-Agents — Track spawned background sessions
• 📊 System Metrics — Uptime, memory pressure, load average

Quick Start

Prerequisites

• Node.js 18+
• Rust (for Tauri)
• macOS (uses launchd, CodexBar)

Install & Run

# Install frontend dependencies
npm install

# Install server dependencies
cd server && npm install && cd ..

# Start API server (port 3001)
node server/api.js &

# Run in dev mode
npm run tauri:dev

Build for Production

npm run tauri:build

The app bundle will be in src-tauri/target/release/bundle/.

API Server

The dashboard requires a local API server that:

• Scans OpenClaw session logs for usage data
• Watches session transcripts for auto-activity detection
• Exposes SSE endpoint for real-time updates

# Log activity from scripts
curl -X POST localhost:3001/api/activity \
  -H "Content-Type: application/json" \
  -d '{"agent":"REM","message":"Task done","type":"success"}'

Configuration

The dashboard reads from:

• ~/.openclaw/agents/main/sessions/*.jsonl — Session transcripts
• CodexBar CLI — Claude Max usage data
• launchd — Cron job schedules

Project Structure

rem-dashboard/
├── src/                    # React frontend
│   └── rem-dashboard.jsx   # Main dashboard component
├── src-tauri/              # Rust/Tauri backend
├── server/                 # Node.js API server
│   ├── api.js              # Express server + SSE
│   └── usage-scanner.js    # OpenClaw log parser
└── start-dashboard.sh      # Launch script

License

MIT