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

Setting it up right now? The Coil × Robinhood pairing recipe is the short version: two MCP servers in one client, the read order that makes it work, and a copy-paste prompt that encodes the discipline.

Coil is a Python engine that runs on Mac or Windows — the agent-led onboarding detects your OS and branches the install for it. 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. Robinhood officially supports agentic trading — dedicated agentic accounts an AI agent connects to over MCP to place trades — so the rails are the broker's own product, not a workaround.
• The data feed — a free Alpaca market-data key for prices and history (data only — no funding, no brokerage there). Cold-backtesting runs against this too.

Robinhood is explicit that it "does not control, supervise, monitor, recommend, or audit these AI agents." That is exactly why the strategy on top has to carry its own spine: Robinhood provides the official rails; Coil provides the validated strategy and the safety core — account lockdown, position caps, drawdown circuit breakers, and a readable audit trail — so what runs on those rails is a discipline you can inspect, not a black box.

It trades the market's leaders, not one pair. The scanner scores every name across the S&P 500, the Nasdaq-100, and a Macro book (bonds, income, gold and metals, commodities) for opportunity, entry-window, hold-conviction, leadership, and sector-rotation phase — then the long-only engine buys the names scored READY or SETUP at real entry windows. It refuses the two ways people lose: it never buys FALLING names (falling knives) and never CHASEs extended names pinned at their highs. When a leader is worth accelerating, it maps to a leveraged vehicle at reduced notional (for example NVDA→NVDL, or a leveraged semis vehicle when semis lead) rather than oversizing the underlying. Every position carries structural stops and ladder (scale-out) exits with trailing stops; when nothing qualifies, cash is a position and the pool sits in the broker's sweep. For the leverage mechanics that make those accelerator vehicles 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 about an hour from download to a scheduled dry-run. (For the broker side — opening the agentic account and connecting your agent over MCP, before Coil enters the picture at all — there's a dedicated step-by-step: how to set up Robinhood agentic trading with Claude Code.) Roughly:

1. Download the engine ($29, one-time), unzip it on your Mac or Windows machine, and point your agent at the folder.
2. Let the interactive onboarding detect your OS, install the Python dependencies, and wire a free Alpaca data key — minutes to a first scan and dashboard.
3. Run the shipped research backtest to see the leadership-rotation backbone replayed on real tape, and the safety self-test suite to verify the spine.
4. Let it run in dry-run, watching, while you learn its behaviour.
5. Going live is a separate, deliberately gated flow: your own account allowlisted, an integrity re-pin, a green safety self-test, and a typed total-loss acknowledgment. It ships disarmed.

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

The honesty layer: the research backbone

Coil ships with a safety self-test suite you can run yourself — the account whitelist, the hard caps, the drawdown ladder, the kill switch — plus the same research-backtest harness that produced the numbers below. The validated rule-set is frozen and fingerprinted: Coil is not an open-ended AI inventing new trades on your machine. Nightly, a small local loop retunes position sizing only, to your own fills, within hard bounds — it never touches entries, exits, stops, or caps. New validated structure arrives only as a deliberate new version you download when you want it. That's the brand — the honest one.

The research backbone (2017–2026 H1)

These are research-backtest figures — not live results, not client returns, and not a prediction. The leadership-rotation backbone Coil's scoring is built on was replayed point-in-time and survivorship-free (delisted names included), with next-open fills and costs modeled. Every number rides next to its same-window SPY benchmark.

In that research backtest the backbone compounded +638% versus SPY's +282%, with a shallower worst drawdown (−23% vs −32%) and a positive result in 9 of 10 years (worst −1%, in 2018). Read the honest rider with it: through the end of 2025 it ran roughly even with SPY at about one-third less drawdown — the outperformance concentrates in leadership regimes (2025 +51%, 2026 H1 +86%). The engine itself is newly live; these are research figures, not a live track record, and past performance does not predict future results.

The risks, stated plainly

• Leverage cuts both ways. The leveraged ETFs Coil uses to accelerate a leader move roughly 2–3x the underlying before gaps and slippage; they decay on multi-day holds and can lose rapidly, up to total loss. Coil uses them at reduced notional for that reason — but reduced size is not no risk.
• 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.

Common questions from operators

Why does my AI agent refuse to place a trade?

In almost every case this is correct behaviour, not a bug. Work down this list in order — each cause looks identical from the chat window ("it just won't trade"), and each has a different fix:

1. No authorised path to your brokerage. A general-purpose assistant cannot place an order because nothing has connected it to one, and inventing a way to move your money would be exactly the wrong instinct. If you have not added the broker's MCP server to your agent, this is your case — see the setup guide, including the --transport http flag most people miss.
2. The MCP is connected, but the account isn't agentic. Robinhood's place_equity_order requires an account with agentic trading enabled and rejects orders from any other account — so the tools all appear, every read works, and every order is refused. Enable agentic trading on the specific account you named, in the Robinhood app. (Verified against Robinhood's Trading MCP as of July 2026; their docs are authoritative.)
3. Nobody was there to confirm. The documented order flow is review_equity_order → present the estimate and alerts → explicit user confirmation → place_equity_order. A scheduled, unattended run with nobody present to confirm is supposed to place nothing on that path — that is the broker's design working. An unattended routine needs its own deliberate, pre-authorised rules for what it may do (which is exactly what an engine like Coil's arming flow is for), not a bypass of the confirmation.
4. If you are running Coil: it ships deliberately disarmed. Live trading is off by default. The engine scans, scores every name, shows you precisely the plan it would execute — and places nothing — until you complete the arming flow yourself.
5. You asked for a decision, not an execution. "Should I buy NVDA?" asks the model to originate an investment recommendation, and it will decline — that refusal is correct and is not a configuration problem. Rules you wrote, executed by rule, are a different request from advice.

The fix is not prompt-wrangling. An agent that can be talked into placing orders it was never configured to place is a liability, not a feature: the same flexibility that lets you argue it into a trade lets a bad instruction do the same. What you want is a real, authorised broker connection plus a deliberate arming step that you perform knowingly.

Can Claude actually execute trades on Robinhood?

Yes — through a broker connector, not by asking nicely. Coil is built for Claude Code and places orders through Robinhood's MCP connector (any broker with an equivalent MCP works). Your credentials stay on your own machine; Coil never collects or transmits them, and it never sees your balances or positions on any server of ours.

How do I let the agent place real orders?

Arming is deliberately a separate, several-step act rather than a toggle: you allowlist the single account it is permitted to touch, run an integrity re-pin, pass a green safety self-test, and type an explicit acknowledgment that you can lose money. Only then does it trade — and a single command stops it again. The friction is the point. Anything easier would be easier to do by accident.

Is it safe to let an AI agent trade my account?

Safety here is bounded, not guaranteed, and we would rather say so plainly. Coil restricts itself to one allowlisted account, enforces hard position and exposure caps, runs a drawdown breaker tied to your equity high-water mark, and fails closed — any ambiguity and it does nothing rather than guess. Those bounds reduce and cap damage. They do not remove market risk, and you can still lose money.

Does Robinhood agentic trading work with Claude Code?

Yes. Robinhood's Trading MCP server is a standard remote MCP endpoint, so any MCP-capable client can use it — Claude Code, Claude Desktop, Cursor, or another host. Claude Code is the better fit for an unattended routine, because it can run on a schedule and execute code between tool calls; Claude Desktop suits a human-in-the-loop session where you approve each step.

Coil's own board is also an MCP server, so both halves load into the same client:

claude mcp add --transport http coil https://coil.trade/mcp

The two-server pairing recipe shows the read order — regime first, then the ranked list, then the broker's tools last.

Can I connect Robinhood agentic trading to Gemini CLI, Codex, or Antigravity?

Yes, in the sense that matters: Robinhood's endpoint is a standard remote MCP server, so this is a config-format question, not a compatibility question. Any client that speaks remote (streamable-HTTP) MCP can connect. Where each one puts it, as of July 2026 — the vendor's own MCP docs are authoritative and these are theirs to change:

The URL itself comes from Robinhood's agentic-trading page once you've enabled it — we deliberately don't print it. One honesty note: Coil is built and tested for Claude Code; other agents can read the same board and the same broker MCP, but we make no promises about how well they run an unattended routine.

What prompt or parameters should I give Claude for agentic trading?

The prompt that matters is not a magic phrase — it is a read order plus your own risk limits, stated explicitly. An agent told only "trade my account profitably" will screen whatever it happened to see last; an agent told to check the market regime first, take candidates only from a ranked list, and size from rules you wrote will behave consistently.

A copy-paste prompt that encodes exactly that discipline is on the Coil × Robinhood pairing page, along with the read order it depends on. Two things belong in your half of it and never the publisher's: position size and exit rules. Coil publishes scores and states — never position sizes, stop prices or target prices.

Why can't the agent place a trailing stop on Robinhood?

Because the order type isn't there — and the workaround is better than the order type. As of July 2026, Robinhood's Trading MCP place_equity_order accepts market, limit, stop_market and stop_limit, with a stop_price and good-til-cancelled support. What it does not accept is the trailing variant. Two more constraints worth knowing before you design around it: market and stop orders are regular-hours only (they're rejected if tagged to an extended session), and the overnight session executes limit orders only — so a resting stop cannot even be parked overnight. Robinhood's documentation is authoritative on all of this and it is theirs to change.

The honest reframe: a trailing stop is arithmetic, not an order type. A scheduled agent recomputes its exit level from current structure each session and sends a plain order when the condition trips — which is exactly how Coil is built: deterministic per-session exits, no resting stop at the broker (see trailing stop for the mechanic). One caveat no order type escapes: nothing defends a price the market gapped over, whether the instruction lives at the broker or in your agent.

Can I have it day-trade for me?

By rule, in the tape that calls for it — not as a scalping alert stream. Coil runs two horizons and lets the market pick: above the long-term trend line it swings the leaders, holding for days; below it, it shortens its horizon and trades session leaders intraday by rule, because in a falling tape overnight exposure is the risk. Either way a name has to prove it's ready — score a real entry — before anything is bought. What Coil is not: a scalper, an open-fader, or an intraday alert stream you react to (the Trade Ideas comparison makes that case in full). The free board is one market day delayed — do not wire an unattended intraday loop against delayed data; the live intraday board is the $12/mo Scanner. Your broker's own day-trade limits apply to your account regardless of what your prompt says.

Which Robinhood MCP tools does an agent actually call, and can it run on a free account?

Robinhood's Trading MCP exposes read tools alongside the order tools, so a well-built routine spends most of its calls reading and places orders rarely. The equity subset an equities routine actually calls, as of July 2026 (verified against the live server — Robinhood's own documentation is authoritative and may change; options, watchlist, scanner and earnings tools also exist beyond this list):

Two names people search for that do not exist: get_account (it's get_accounts) and get_orders (it's get_equity_orders).

What we can say about our side: Coil's board has a free tier with no key, wallet or account — the full board one market day delayed, which is enough for a swing-timeframe agent. There is no time limit on it and nothing expires. On session length: a swing routine is a handful of tool calls a few times a day — read the regime, read the ranked list, check positions, maybe place an order — which is why it fits comfortably inside a modest agent plan. What burns through a plan's limits is an always-on intraday watcher polling every few minutes; that is a different product shape, and if you want it, budget for it deliberately. Account eligibility on the broker side is governed by Robinhood, not by Coil — the setup guide's cost section breaks down all three layers.

Pricing and next step

Coil is $29 one-time (regular $49) for the full engine — scanner, dashboard, and the long-only opportunity engine — yours to keep and run forever. No subscription, no tiers, no recurring charge. New versions are announced on coil.trade and installed manually when you want them. Checkout runs through Gumroad as merchant of record — it handles checkout, receipts, and license keys. Because Coil is an instant download you keep, all sales are final; the only thing corrected is a genuine billing error, such as a duplicate charge. See pricing, or read how it works first if you'd rather see the engine before you decide.