🦞 LobsterMC — AI Trading Command Center

A mission control dashboard for AI trading agents. Built for quant traders who run their own AI company.

What is this?

LobsterMC is a real-time command center that visualizes your AI trading team, live signals, portfolio positions, and automation schedules — all in one dark-mode dashboard.

It's built for people who:

• Run multiple AI agents for market research and trading
• Want a Bloomberg-style interface for their quant system
• Need to monitor signals, positions, and agent activity at a glance

Features

• 📡 Live Signal Panel — Critical/High severity signals with 3-layer resonance analysis
• 📊 Portfolio Tracking — Real-time P&L, positions, cost basis, signal scores
• 🤖 AI Company — Visualize your agent team (roles, models, online status)
• 🧠 Strategy Engine — 3-layer system: Early Warning → Macro Trigger → AlphaWhisper
• ⚡ Activity Feed — Real-time log of agent actions and decisions
• 🕐 Cron Monitor — View and track all scheduled automation jobs
• 💰 Token Usage — Cost breakdown by model and agent
• 📋 Task Queue — Completed and pending task tracking
• Bloomberg Ticker — Live scrolling market data at the bottom

Screenshots

Stack

• Frontend: Pure HTML + CSS + Vanilla JS (single file, zero dependencies)
• Backend: Python Flask
• Data: Reads local JSON files (portfolio, signals, agent state)
• Design: Linear/Vercel-inspired dark UI, particle canvas background

Quick Start

git clone https://github.com/1m1ai/LobsterMC
cd LobsterMC/backend
pip install flask
python server.py
# Open http://127.0.0.1:19001

Architecture

LobsterMC/
├── frontend/
│   └── index.html          # Entire UI (single file)
├── backend/
│   └── server.py           # Flask API server (port 19001)
└── docs/
    └── preview.png

API

The backend exposes one endpoint:

GET /api/status

Returns portfolio, active signals, execution plans, and agent state. Connect your own data sources by editing server.py.

Data Sources

By default reads from:

• paper_trading/portfolio.json — positions and P&L
• paper_trading/monday_plan.json — signals and execution plans
• Star-Office-UI/state.json — agent status

Swap these out for your own data pipeline.

The 3-Layer Strategy System

Layer 1: Early Warning (Smart Money)
  → Identify signals 3 months before events
  → Entry at 5% position

Layer 2: Macro Trigger
  → 25 global trigger conditions
  → Entry at 10% position

Layer 3: AlphaWhisper
  → Intercept retail platform signals before the crowd
  → Cooperation keywords: 72~78% win rate

All 3 layers active → Full position (20%)

Inspired By

• mission-control — great concept, we went further
• Bloomberg Terminal — the original trading UI
• Linear, Vercel — modern dark design language

License

MIT

---

Built by an AI trading team. For AI trading teams.