This week the automated trading stack went fully live — three independent bots across three different markets, all feeding a single dashboard that gives us a real-time view of where capital stands across the whole portfolio. Here's what that actually looks like in practice.
// What the stack actually doesWe run three trading bots simultaneously, each handling a different market and strategy:
Each bot has its own timer, its own risk rules, and its own dashboard panel. They share nothing except the consolidated finance view.
// Why we built it this wayThree bots, three markets, zero mixing. That's intentional. A crypto bot and a stock trader have completely different risk profiles, different order types, and different failure modes. Blending them in one system means you can't kill one without killing the other, and you can't reason about them separately.
Separation also means honesty. When Alpaca reports its P&L, that's Alpaca's P&L — not a crypto position bleeding into the equity report, or vice versa. The consolidated dashboard adds them up for the big picture, but the source of truth for each lane stays in its own lane.
The bots are fully automated within their scope. Configuration changes, stale-order cleanup, and parameter tuning happen without a manual gate. We escalate only when capital is being added, stop-loss policy changes, or something irreversible is on the table.
// How it fits into the labThese aren't speculative projects sitting in a repo. They're running on dedicated infrastructure — a Linux VM that keeps the timers firing around the clock, with the Alpaca MCP running as a sidecar for read-only API visibility. The dashboard is internal only, pulling live numbers from each system's API rather than from a local spreadsheet.
This is the same operational discipline we apply to everything in the lab: verified, monitored, and wired into the alert layer. If a timer misses a beat, we know. If a bot's equity drifts outside expected bounds, we know. The goal isn't just to run the bots — it's to run them reliably, over months, without babysitting.
The Alpaca live account has real capital deployed. The Kalshi and crypto lanes are tracking live market data even in paper mode. Everything is wired, everything is ticking, and the dashboard reflects what's actually happening — not what we hope is happening.
When the crypto bot validates, it graduates to live capital the same way the equities bot did: with the same risk rules, the same separation, and the same "nothing irreversible without a human decision" discipline.
// A small note on transparencyWe're not going to announce every trade. The point of automating is to remove emotion from execution, and publishing a play-by-play defeats that. What we will do is report honestly: monthly realized P&L, significant drawdowns, and any material changes to how the strategies work. The dashboards are for us — the public reporting is for accountability.
Questions about the stack? Reach us at contact@graveystudios.com.