How to run an AI agent that trades on Robinhood

You want your AI agent to place trades on Robinhood for you. The short, honest answer: an AI agent can run a rules-based trading program on a schedule — reading the market, deciding within fixed rules, and routing orders through your broker — but it should never be a black box you hand your money to. The good versions are software you operate, on your machine, with your broker and your capital. Coil is one of those. Here's how it works and what's realistic.

What "agentic trading" actually means

"Agentic" just means an AI model can take actions through tools, not only chat. In trading, that looks like a scheduled session waking up, pulling fresh market data, evaluating a fixed set of rules, and — if and only if the rules say so — placing an order through a broker connection. The intelligence is in the operating: running the same disciplined checklist every session without getting bored, scared, or greedy.

The dangerous version is an LLM improvising trades from a vague prompt. The sound version is the opposite: the agent executes a deterministic, pre-written strategy and the model's judgement is boxed in by hard rules it cannot override. Coil is built the second way.

How Coil wires Claude to Robinhood

Coil is a Python engine that runs on macOS. Three pieces connect:

• The operator — scheduled Claude sessions. At set times, a Claude session runs the engine: fetch data, score the setup, decide, and either act or stand down. Built for Claude specifically.
• The broker connector — built for Robinhood via an MCP (a standard tool interface), and it works with any equivalent broker MCP. This is the layer that actually reads positions and places orders.
• The data feed — a free Alpaca market-data key for prices and history. Cold-backtesting runs against this too.

It trades exactly one pair: SOXL (3x long semiconductors) and SOXS (3x inverse). Its working hypothesis is a "compression-to-ignition" pattern — the idea that semis tighten into low-volatility coils and then tend to expand — which the engine treats as a setup to test, not a market truth to bet on blindly, paired with leg-rider entries and rule-based exits. Whether the pattern holds in any given week is exactly what the cold harness and live probation are there to check. For the leverage mechanics that make this pair both potent and punishing, read SOXL and leveraged-ETF decay.

What you keep control of

This is the part that separates "software you run" from "handing over your money."

• Your keys never leave your machine. Broker and data credentials stay local. Coil does not route your account through a server it owns.
• Your capital, your account. You fund and watch your own Robinhood account. There's no pooled money and no custody transfer.
• You set the schedule and can stop it. Sessions are scheduled by you; you can pause, change, or kill them anytime. Standing down is a position — when rules say "no trade," idle cash sits in Robinhood's cash sweep (around 3.35% APY, variable, on Robinhood Gold, as of Feb 2026 — that's the broker's yield, not Coil's, and not risk-free).
• You're the overnight backstop. Coil uses deterministic per-session exits and account-level circuit-breakers, but a human watching is always the strongest safety layer.

The realistic setup

Budget one to two hours for first-time setup. Roughly:

1. Download the engine ($5, one-time) and unzip it on a Mac.
2. Install Python dependencies and a free Alpaca data key.
3. Connect your broker MCP (Robinhood or equivalent) with your own credentials.
4. Run the cold-backtest harness yourself to see the logic before risking a cent.
5. Schedule the Claude sessions, ideally starting tiny while you learn its behaviour.

For a deeper walkthrough of the engine's internals and the validation harness, see how it works.

The honesty layer: cold validation

Coil ships with a backtest A/B harness running 20 integrity guards, plus two local self-tuning loops that re-fit only within hard whitelisted bounds and auto-rollback if a change underperforms. "Cold validation" means a fresh process per market regime — no cross-run cache warmth flattering the numbers. The harness is rigorous enough that it caught three of its own backtest bugs: a look-ahead, a next-day leak, and a sign-inverted short book. That's the brand — the honest one.

Backtested regime results

These are backtested, best-case figures — not typical, not a prediction, and not client returns. Treat them as hypotheses to test yourself, not as anything you should expect to repeat.

Three of four backtested regimes are profitable; the bear is the honest weak spot — its profit factor is still below 1.0 — and it now stands down to cash on confirmed sustained-bear days. The sample is thin: roughly 115 trades a year on one ETF pair, under ~500 trades total. Small samples can flatter or fool. Live, forward results on probation matter more than any backtest.

The risks, stated plainly

• Leverage cuts both ways. A ~10% move in the semis index is roughly ~30% in these ETFs before gaps and slippage. They decay on multi-day holds and can lose rapidly, up to total loss.
• Circuit-breakers reduce damage; they don't remove risk. Coil's breakers tie to your account's high-water mark to cap single-day pain — but stops can gap through on a hard open.
• This is not passive income. No guaranteed return, no autopilot wealth. You're operating risk software, and you can lose money.

If you're weighing this against off-the-shelf bots and paid alerts, the Coil vs trading bots and signal services comparison lays out the differences honestly.

Pricing and next step

Coil is $5 one-time for the full self-tuning engine, yours to keep. Pro is $25/mo or $249/yr and adds Coil Sync — each new cold-validated version delivered autonomously. Checkout runs through Lemon Squeezy (merchant of record, powered by Stripe). See pricing, or read how it works first if you'd rather see the engine before you decide.