tenacitOS

TenacitOS — Mission Control

A real-time dashboard and control center for OpenClaw AI agent instances. Built with Next.js, React 19, and Tailwind CSS v4.

TenacitOS lives inside your OpenClaw workspace and reads its configuration, agents, sessions, memory, and logs directly from the host. No extra database or backend required — OpenClaw is the backend.

---

Features

• ** System Monitor** — Real-time VPS metrics (CPU, RAM, Disk, Network) + PM2/Docker status
• ** Agent Dashboard** — All agents, their sessions, token usage, model, and activity status
• ** Cost Tracking** — Real cost analytics from OpenClaw sessions (SQLite)
• ⏰ Cron Manager — Visual cron manager with weekly timeline, run history, and manual triggers
• ** Activity Feed** — Real-time log of agent actions with heatmap and charts
• ** Memory Browser** — Explore, search, and edit agent memory files
• ** File Browser** — Navigate workspace files with preview and in-browser editing
• ** Global Search** — Full-text search across memory and workspace files
• ** Notifications** — Real-time notification center with unread badge
• ** Office 3D** — Interactive 3D office with one desk per agent (React Three Fiber)
• ** Terminal** — Read-only terminal for safe status commands
• ** Auth** — Password-protected with rate limiting and secure cookie

---

Screenshots

Dashboard — activity overview, agent status, and weather widget

Session History — all OpenClaw sessions with token usage and context tracking

Costs & Analytics — daily cost trends and breakdown per agent

System Monitor — real-time CPU, RAM, Disk, and Network metrics

Office 3D — interactive 3D office with one voxel avatar per agent (React Three Fiber)

---

Requirements

• Node.js 18+ (tested with v22)
• OpenClaw installed and running on the same host
• PM2 or systemd (recommended for production)
• Caddy or another reverse proxy (for HTTPS in production)

---

How it works

TenacitOS reads directly from your OpenClaw installation:

/root/.openclaw/              ← OPENCLAW_DIR (configurable)
├── openclaw.json             ← agents list, channels, models config
├── workspace/                ← main agent workspace (MEMORY.md, SOUL.md, etc.)
├── workspace-studio/         ← sub-agent workspaces
├── workspace-infra/
├── ...
└── workspace/mission-control/ ← TenacitOS lives here

The app uses OPENCLAW_DIR to locate openclaw.json and all workspaces. No manual agent configuration needed — agents are auto-discovered from openclaw.json.

---

Installation

1. Clone into your OpenClaw workspace

cd /root/.openclaw/workspace   # or your OPENCLAW_DIR/workspace
git clone https://github.com/carlosazaustre/tenacitOS.git mission-control
cd mission-control
npm install

2. Configure environment

cp .env.example .env.local

Edit .env.local:

…

Tip: OPENCLAW_DIR defaults to /root/.openclaw. If your OpenClaw is installed elsewhere, set this variable.

3. Initialize data files

cp data/cron-jobs.example.json data/cron-jobs.json
cp data/activities.example.json data/activities.json
cp data/notifications.example.json data/notifications.json
cp data/configured-skills.example.json data/configured-skills.json
cp data/tasks.example.json data/tasks.json

4. Generate secrets

# Auth secret
openssl rand -base64 32

# Password (or use a password manager)
openssl rand -base64 18

5. Run

# Development
npm run dev
# → http://localhost:3000

# Production build
npm run build
npm start

Login at http://localhost:3000 with the ADMIN_PASSWORD you set.

---

Production Deployment

PM2 (recommended)

npm run build

pm2 start npm --name "mission-control" -- start
pm2 save
pm2 startup   # enable auto-restart on reboot

systemd

Create /etc/systemd/system/mission-control.service:

[Unit]
Description=TenacitOS — OpenClaw Mission Control
After=network.target

[Service]
Type=simple
User=root
WorkingDirectory=/root/.openclaw/workspace/mission-control
ExecStart=/usr/bin/npm start
Restart=always
RestartSec=10
Environment=NODE_ENV=production

[Install]
WantedBy=multi-user.target

sudo systemctl daemon-reload
sudo systemctl enable mission-control
sudo systemctl start mission-control

Reverse proxy — Caddy (HTTPS)

mission-control.yourdomain.com {
    reverse_proxy localhost:3000
}

When behind HTTPS, secure: true is set automatically on the auth cookie.

---

Configuration

Agent branding

All personal data stays in .env.local (gitignored). The src/config/branding.ts file reads from env vars — never edit it directly with your personal data.

Agent discovery

Agents are auto-discovered from openclaw.json at startup. The /api/agents endpoint reads:

{
  "agents": {
    "list": [
      { "id": "main", "name": "...", "workspace": "...", "model": {...} },
      { "id": "studio", "name": "...", "workspace": "..." }
    ]
  }
}

Each agent can define its own visual appearance in openclaw.json:

{
  "id": "studio",
  "name": "My Studio Agent",
  "ui": {
    "emoji": "",
    "color": "#E91E63"
  }
}

Office 3D — agent positions

The 3D office has default positions for up to 6 agents. To customize positions, names, and colors for your own agents, edit src/components/Office3D/agentsConfig.ts:

export const AGENTS: AgentConfig[] = [
  {
    id: "main",       // must match workspace ID
    name: "...",      // display name (can also come from API)
    emoji: "",
    position: [0, 0, 0],
    color: "#FFCC00",
    role: "Main Agent",
  },
  // add your sub-agents here
];

3D Avatar models

To add custom 3D avatars (Ready Player Me GLB format), place them in public/models/:

public/models/
├── main.glb        ← main agent avatar
├── studio.glb      ← workspace-studio agent
└── infra.glb       ← workspace-infra agent

Filename must match the agent id. If no file is found, a colored sphere is shown as fallback.
See public/models/README.md for full instructions.

Cost tracking

Usage is collected from OpenClaw's SQLite databases via a script:

# Collect once
npx tsx scripts/collect-usage.ts

# Auto-collect every hour (adds a cron job)
./scripts/setup-cron.sh

See docs/COST-TRACKING.md for details.

---

Project Structure

…

---

Security

• All routes (including all /api/*) require authentication — handled by src/middleware.ts
• /api/auth/login and /api/health are the only public endpoints
• Login is rate-limited: 5 failed attempts → 15-minute lockout per IP
• Auth cookie is httpOnly, sameSite: lax, and secure in production
• Terminal API uses a strict command allowlist — env, curl, wget, node, python are blocked
• Never commit .env.local — it contains your credentials

Generate fresh secrets:

openssl rand -base64 32   # AUTH_SECRET
openssl rand -base64 18   # ADMIN_PASSWORD

---

Troubleshooting

"Gateway not reachable" / agent data missing

openclaw status
openclaw gateway start   # if not running

"Database not found" (cost tracking)

npx tsx scripts/collect-usage.ts

Build errors after pulling updates

rm -rf .next node_modules
npm install
npm run build

Scripts not executable

chmod +x scripts/*.sh

---

Tech Stack

Layer	Tech
Framework	Next.js 15 (App Router)
UI	React 19 + Tailwind CSS v4
3D	React Three Fiber + Drei
Charts	Recharts
Icons	Lucide React
Database	SQLite (better-sqlite3)
Runtime	Node.js 22

---

Contributing

1. Fork the repo
2. Create a feature branch (git checkout -b feat/my-feature)
3. Keep personal data out of commits — use .env.local and data/ (both gitignored)
4. Write clear commit messages
5. Open a PR

See CONTRIBUTING.md for more details.

---

License

MIT — see LICENSE

---

Links

• OpenClaw — the AI agent runtime this dashboard is built for
• OpenClaw Docs
• Discord Community
• GitHub Issues — bug reports and feature requests