Connect an Alpaca paper account and you get a full brokerage environment with simulated money — real market data, real order types, real charts, nothing at risk. What makes it a learning tool rather than a practice tool is the assistant inside it: it draws, it analyses, it explains its reasoning, and it cannot place a trade.
Educational. Not investment advice.
An Alpaca paper account gives you a funded simulated brokerage: stocks and crypto, live market data, market and limit orders, positions, fills, and a P&L that moves. No money involved.
Ask it to draw support and resistance and it draws them — then tells you how many times each level was actually touched. Not “strong support,” but touched four times, most recently two days ago.
Ask why a setup looks weak and it walks the reasoning out loud: structure is a downtrend, momentum is rising, that’s countertrend conflict, and here’s what would need to happen to confirm otherwise. You pick up the vocabulary by watching it used correctly on your own chart.
These are real exchanges from the terminal, replayed. Click a question and watch what the assistant does — where it acts, where it asks first, and where it stops.
Ask in plain language and it acts on the chart you named:
With four charts open, a request about “the ETH chart” lands on that one only. If it’s ambiguous, it asks by number, venue and timeframe rather than guessing.
Everything it can draw, you can draw yourself — full drawing tools and indicators with editable parameters. Your charts, timeframes, indicators and hand-drawn overlays all persist. Close the tab, come back tomorrow, the workspace is where you left it.
This is architectural, not a setting.
Learning to trade alongside a tool that will say “I can’t see that” is worth considerably more than one that always has an answer.
Paper trading teaches mechanics, not psychology. Fills are simulated, so slippage on thin instruments won’t feel realistic. And risking nothing feels nothing like risking something — the hardest part of trading is the part a simulator cannot reproduce.
Use it for what it’s genuinely good at: learning the interface, learning what indicators actually measure, testing whether an idea survives contact with real price data, and building the habit of checking before acting. Then size down hard when you move to real money.
Real-money crypto through Coinbase is supported. It’s off by default, it requires an explicit setting, and it’s bounded by a maximum order size you set yourself — a limit the server enforces, not the interface.
That second list isn’t a roadmap. Those are deliberate limits, and they’re staying.
A free Alpaca paper account, which takes a few minutes and involves no money.
Not on a paper account. Real-money trading is a separate connection that’s off by default and capped when you turn it on.
Alpaca for stocks and crypto, paper and live. Coinbase for real-money crypto.
No. It explains what the terminal is showing and reasons about it out loud. It doesn’t tell you what to buy, size positions, or predict prices — and it isn’t a licensed advisor.
No. One subscription covers everything. See pricing ›
The assistant has been put through repeated scored reviews by an external trader whose job was to break it — verifying every figure against what the terminal actually published, and separating “the terminal never sent this” from “the assistant reasoned badly.” Those look identical in the output, and only one of them is the assistant’s fault.
Most of what that found were not hallucinations. They were missing fields — places where the terminal was silent and the model filled the gap. Each one got fixed by publishing the fact rather than by adding a rule telling the model to behave. Turning a rule into a computed fact changes behaviour where a rule does not.
Beta testers get first access — and founding-member pricing when we launch. No card required. See pricing ›
We’ll only email you about beta access and launch. Unsubscribe anytime.