Alpaca Paper Rails
A real brokerage API in simulation mode: $1M paper equity, live market prints, real order plumbing. The engine practices on the same tracks it will race on.
A real brokerage API in simulation mode: $1M paper equity, live market prints, real order plumbing. The engine practices on the same tracks it will race on.
One gatekeeper between the web and the account. Keys live server-side only — no browser ever sees a credential. Every price, every headline, every position flows through it.
The command view: live tape, 36 stocks, sectors, metals, the news wire, account and positions — Bloomberg-caliber density, Wilder-caliber taste.
The engine is built and audited by a crew of AI seats — GLT builds, MMT ships, KKT designs, GPT organizes, under human command. Every build carries receipts; every save carries a spoken GO.
Data-dense mission control: multi-timeframe monitors, options-flow telemetry, a velocity race widget ranking live momentum, agent execution logs streaming in real time. Modeled on the best of Trade Ideas and TrendSpider — but ours.
The other half of the terminal: one omni-prompt bar, plain-English command of the book — "what's our tech exposure," "hedge the winners" — with agent reasoning cards explaining every move in human words before anything executes.
Agents that scan, propose, and wait. Every trade surfaces as a confirmation card — approve or abort — until the desk has earned the right to let the leash out. Trust is built on receipts, one session at a time.
"The goal isn't a robot that trades. It's a crew that never gets tired, never gets greedy, and never touches the vault."