Trade Robinhood Chain like a pro.
Factor is a native macOS + iOS terminal for Robinhood Chain: a portfolio, a live signal board across tokenized stocks, autonomous paper agents that run around the clock, non-custodial swaps, and a token launcher. This guide is everything you need to use it.
What is Factor
Factor is a non-custodial terminal for trading tokenized real-world assets — tokenized stocks and ETFs — on Robinhood Chain. It is not a broker or an exchange: it never holds your keys or your funds. It reads the chain, computes signals, and — when you choose to go live — routes trades that your own wallet signs.
A read cockpit
Your portfolio, a blended signal board across tokenized stocks, and a per-token thesis explaining why a name looks bullish or bearish.
Paper & agents
A free manual paper account, plus named agents that trade a basket and book their own P&L — running 24/7 even with the app closed.
Execution
Non-custodial swaps and a token launcher, signed by your wallet. Live trading stays behind a deliberate arm + fund step.
An on-device advisor
Apple Intelligence narrates why a trade looks strong or risky — grounded in the signals. It explains; it never decides for you.
Factor explains and routes; you decide and sign. The advisor explains, the agents trade paper, and nothing touches real capital until you arm it and fund it yourself.
Is it safe?
Factor is non-custodial by design. There is no step where Factor holds your private key or moves your funds on its own.
- Your wallet signs, always. Swaps and payments are signed by your own wallet in a secure system browser sheet. Factor only receives a confirmation that a transaction happened.
- Live trading is opt-in and gated. Agents start in paper. Going live for real capital takes a deliberate confirmation, and you fund it — Factor can't move money you haven't put behind it.
- Paper needs no wallet at all. You can use the whole terminal and run agents in paper with no wallet connected.
- Paper is sandboxed per device and per account. Trying Factor without signing in gives you your own private paper account, not a shared one. Sign in later and it follows you.
- Live trading needs a verified sign-in and an active membership. A demo or try-it session is paper-only by construction — it cannot arm an agent, whatever else it can do.
- Adversarially tested. The security-critical paths (sign-in, entitlements, payment verification) are covered by dedicated test suites, and payments are re-checked on-chain rather than trusted from the app. The tests are written against the failure this product cares about most: a number that looks real but isn't.
- External audit: pending. The on-chain contracts have a full test suite and every security-critical guard is mutation-verified — each one has been deliberately broken to confirm a test catches it. They have not yet been reviewed by an outside firm. We will name the firm and link the report here when that is done, and not before.
Non-custodial by design — every trade is signed by your own wallet.
Demo vs live
Factor has two independent switches, and it helps to keep them straight:
| Switch | What it changes |
|---|---|
| Data demo ↔ live | Whether you're exploring with a sample account or connected to your real holdings. Even in demo, prices and signals are real. |
| Trading paper ↔ live | Whether trades are simulated (paper — free, no risk) or signed on-chain by your wallet (live). |
You start in demo + paper: a live-feeling terminal with real prices and real signals, simulated fills, and zero risk. No wallet required to look around.
Portfolio
Your home base. In paper mode this is your manual paper account — equity, day change, open positions, and full trade history, marked to live prices. Connect a wallet to see your real on-chain holdings. A hide-balances toggle blanks every dollar figure while keeping the labels and charts, so you can screen-share safely.
Signals & the board
The board (Markets) is one blended view across tokenized stocks and native tokens, every row marked with a signal. Slice it by asset class, by signal direction, or sort by strongest signal / biggest movers.
The Signals screen is a hub for any token. For each, Factor shows a stance — how bullish or bearish the blended signal is — and a confidence in that read, plus a short "thesis" breaking down what's driving it. Factor blends several market factors and, over time, leans on the ones that have actually been predictive and quiets the ones that haven't — so the read improves as evidence accumulates.
Stance = which way and how strongly. Confidence = how much to trust it right now. A strong stance at low confidence is a maybe, not a green light.
The Fund Advisor
On supported devices, Factor's Fund Advisor uses Apple Intelligence, on-device, to narrate a token's thesis in plain language — a verdict, a headline, what's backing it, and what to watch. It's grounded strictly in the signals.
The advisor explains the read; it does not place trades or set the direction. Buy/sell/hold and the numbers come from the deterministic engine. On devices without Apple Intelligence, you still get a clear written summary.
Paper trading
Paper trading is free, unlimited, and needs no wallet. You get one manual paper account that starts with a set balance and marks to live prices. Buy and sell to build positions; every fill is kept in your history; reset any time to start fresh. It's the honest way to try a strategy before a cent is at risk.
Your manual paper account is your own hand-placed trades. Each agent keeps its own separate paper account — so an agent's performance is always attributable to the agent, not mixed in with your manual trades.
Agents
An agent is a named strategy bound to a basket — a curated set of tokenized stocks. It watches its basket around the clock and trades it on the live signals, booking its own P&L in its own separate paper account.
Setting one up is a short conversation: tell Factor what you want to trade in plain language and it assembles the basket and starts the agent. Free includes one agent; membership unlocks more.
Not every tokenized stock can be traded every way. Some have an on-chain spot pool, some have a perpetual market, some have both, and some are signals-only. Factor marks each name in your basket so you can see what an agent could actually act on — a name with no tradeable rail can never produce a trade, however strong its signal.
Setups — what an agent waits for
An agent doesn't buy because a number is high. It waits for a named setup — a specific, recognisable pattern across the signal legs — and it tells you which one fired on every trade it makes.
| Setup | Direction | What it's waiting for |
|---|---|---|
| Trend Continuation | Long / Short | Momentum and the wider regime agree and both are moving. |
| Throughput Breakout | Long | Real on-chain activity is accelerating and order flow confirms it. |
| Mean-Reversion Snapback | Long / Short | Price is stretched away from value in a range, and momentum is fading. |
| Value Dislocation | Long | Trading below fair value with no macro reason for the discount. |
| Liquidity Tailwind | Long | Fresh liquidity building into a risk-on tape. |
| Risk-Off Defense | Short | Macro is risk-off and the regime is deteriorating. |
| Flow Divergence | Long / Short | Order flow turns ahead of price. |
You can switch any setup on or off per agent. A setup whose underlying signals aren't currently available is shown as Unavailable rather than silently counted as "didn't fire" — so you can always tell the difference between the pattern wasn't there and Factor couldn't see.
Because the setup is recorded on the trade, the daily report can tell you which kinds of trade make you money and which lose it — not just your net P&L. That's the difference between a scoreboard and a lesson.
Tuning the strategy
Every agent exposes the dials its engine actually reads. Open Tune on any agent to change them. Nothing is applied until you save, and the panel shows you the exact diff first — leverageCap 3→4 — along with a note field for why you turned it.
| Group | Dials | What they change |
|---|---|---|
| Selection | Entry threshold · Exit threshold · Setups on/off · Signal weights & mutes | How convinced the agent must be to act, and which evidence it listens to. |
| Risk & sizing | Position size · Max positions · Max exposure · Stop distance · Profit target · Leverage cap · Allow shorts | How much it commits and how it protects a position. |
| Regime filters | Volatility gate · Trend gate | Conditions under which it stands down entirely. |
Every change is logged with your rationale and the date. The daily report can then compare how the agent did before and after — so a knob turn becomes a testable decision, not a guess you forget you made.
Each dial has a safe range. If you set something outside it, Factor clamps it and the applied diff shows you the value it actually took. Always read the diff.
Why it didn't trade
A quiet day is not a broken agent, and Factor won't let the two look the same. For every pass over your basket, an agent records the reason each name was passed over:
| Reason | Means |
|---|---|
| Below conviction | The signal wasn't strong enough to clear your entry threshold. |
| No setup | The signal was strong, but it didn't match any named pattern you have switched on. |
| Direction disagreed | The setup and the blended signal pointed opposite ways. |
| Regime gate | Your volatility or trend filter stood the agent down. |
| Already full | The agent is at its maximum number of positions. |
| No budget | Cash or your exposure ceiling left nothing to size a position with. |
| Couldn't act | The setup fired and passed every check, but no rail could express it — for example a short on a name with no perpetual market. |
It is the one reason that isn't about the market. It means the agent wanted to trade and the plumbing wouldn't let it — usually a basket name with no tradeable rail, or shorts switched off. That's a setup change, not a market call.
The daily report
Once a day each agent files a short report on itself. It is deliberately unflattering.
- A scorecard — round-trips completed, hit rate, average result per trade measured in multiples of the risk taken, and a breakdown by setup.
- A baseline — what simply holding the basket would have done. If buy-and-hold beat the agent, the report says so.
- A verdict on quiet days — No-edge sit-out (the agent could see and chose not to trade — a good outcome) is kept separate from Data gap (it couldn't see, so its silence says nothing about edge).
- At most two recommendations, because a list of ten is a list of none. If Factor is partly blind, fixing that outranks any tuning suggestion.
- A no-improvement flag if the agent has failed to beat the boring baseline several days running.
Where a number can't be computed — no completed trades, no prior report to compare against — the report shows nothing rather than a zero. A fabricated 0.00 reads as "flat" when the truth is "unknown," and that is the one thing a trading report must never do.
Chatting with agents
Every agent has a chat button — the speech bubble in the gutter beside its card. Open it to talk to that agent about its own book: ask how it's doing, or tell it to trade, in plain language.
Nothing executes from a message. The chat turns what you say into a proposal card; the trade only happens when you tap Approve. Closing the sidebar or dismissing the card cancels it.
Two rails: equity and perps
Each tokenized stock can trade on one or both rails, shown by a badge on the ticker — and the verb you use picks the rail:
| Badge | Trades on | Say |
|---|---|---|
| E+P | Equity (spot) and perp | buy for spot, or long/short for a perp |
| E | Equity only | buy / sell |
| P | Perp only | long / short (no plain "buy") |
| SIG | Signals only | not tradeable yet |
buy and add mean spot. The words long / short, or any leverage like 2x, mean perp. So buy NVDA and long NVDA are deliberately different orders. Ask to "buy" a perp-only ticker and the chat points you to long/short instead.
What it understands
| You say | It proposes |
|---|---|
buy $500 NVDA | Spot buy |
long NVDA $500 3x · short HOOD perp $500 2x | Open a perp (long / short) |
close TSLA · sell TSLA | Close a position |
reduce NVDA 50% · trim NVDA half | Reduce a position |
close everything · sell all | Flatten the whole book |
sell NVDA at $250 | Limit sell (take-profit) — rests until price rises |
stop NVDA at $200 | Stop-loss — rests until price falls |
Amounts always take a $ ($500, $50k, $1.5m); fractions can be 50%, half, or all; leverage is 2x…10x. You can act on several names at once — Close $HOOD and $NVDA or Buy $250 NVDA and AAPL — up to five cards, each approved on its own.
It's a conversation
You don't have to phrase it perfectly. If something's missing — the side, the rail, an amount, or which position — it asks, and remembers your answer:
You: short HOOD perp Factor: How much margin for the HOOD perp — a dollar amount? You: $500 Factor: Short HOOD perp — $500 at 1×? → card appears
You: trade COIN Factor: COIN trades on both rails — equity (spot) or perp? You: perp Factor: Long or short the COIN perp? You: long $500 3x Factor: Long COIN perp — $500 at 3×? → card appears
A perp-only ticker skips the rail question. Say "never mind" to drop it. If you say close and hold more than one thing, it asks which — answer with the tickers, or "all".
The guardrails
- It won't overspend. Ask for more than the agent's cash and it refuses, tells you the max, and re-asks. A too-big order can't be approved — the card shows ✕ Exceeds cash and the button is disabled.
- Large orders scream. Anything ≥ $50k (or ≥ half the agent's cash) turns the card red, shows the dollars big, and the button reads "Approve anyway." Read the number before you tap.
- It won't trade the wrong rail, and a question is never a trade — "how close am I to breakeven?" is answered, not acted on.
Always put the $ on amounts — it's what separates an amount from a ticker. buy $200 MSTR commits two hundred dollars; the parser will never read the ticker as a multiplier.
24/7 & the debrief
Your agents don't clock out. They keep trading paper on Factor's servers, around the clock, even when the app is closed — so your device isn't doing the work. Open the app and it simply picks up where the agents are.
Once a day (configurable), you get a debrief: a clean report of everything the agents did while you were away — fills, per-agent standings, notable trades. A gentle reminder can nudge you to check in even while the app is closed. You control the cadence, or turn it off, in Settings.
Editing & retiring an agent
Changing the basket. Add or remove names any time. When you remove a name the agent is holding, you choose per position: ride it out (the agent stops entering that name but keeps managing the open position to its stop or target) or close now. Factor never strands a position — a name dropped from the basket is still watched until it's closed.
Retiring. Retiring is a lifecycle, not a delete button. The agent moves to closing, stops taking new positions entirely, and winds down what it holds: perpetuals close immediately since they trade around the clock, equity positions wait for the next market session. Only once it is genuinely flat can it be removed — and if you try sooner, Factor refuses and tells you exactly what is still open.
You can't delete an agent out from under an open position, and you can't lose one to a mis-tap. The refusal carries the reason.
Swaps & going live
Tap Swap anywhere to trade a token. In paper mode you get an honest simulated fill booked into your paper account. In live mode, Factor prepares a real quote and opens a signing sheet where your wallet signs the swap — Factor never holds a key.
Taking an agent from paper to live is deliberate: you confirm the switch and fund it. Until you do, everything stays paper. This arm-and-fund step is the line between "trying it" and "trading it," and it's always your call.
When you go live for the first time, use a tiny amount to confirm everything behaves before sizing up.
Launching a token
Factor includes a token launcher (macOS). Give your token a name, symbol, logo, and links, choose what it pairs against, preview the card, and launch — signed non-custodially by your wallet. It's a secondary tool tucked into the menu, not part of the day-to-day terminal.
Plans & pricing
| Plan | Price | Includes |
|---|---|---|
| Free | $0 | The full read cockpit, live signals preview, unlimited paper trading, and one paper agent. |
| Membership | $119/mo (or $99 on the crypto rail) | Live signals, manual live trading, and the launcher. |
| Extra agents | +$99/mo each (+$79 crypto) | Each additional autonomous agent. |
Annual is roughly two months free. iPhone purchases go through the App Store; on Mac you can pay with crypto for the better price. Paper trading is always free.
Glossary
| Robinhood Chain (RHC) | The chain Factor trades on. |
| Tokenized stock (RWA) | A real-world asset — a stock or ETF — represented as an on-chain token. |
| Stance | How bullish or bearish the blended signal is for a token. |
| Confidence | How much to trust that stance right now. |
| Basket | A curated set of tokens an agent trades. |
| Paper | Simulated trading — free, no wallet, no real capital. |
| Live / armed | Real on-chain trading, signed by your wallet after you confirm and fund it. |
| Non-custodial | Factor never holds your keys; your own wallet signs everything. |
FAQ
Do I need a wallet to try it?
No. The full terminal and paper trading — including agents — work with no wallet. You only connect one to trade live or see real holdings.
Can Factor move my money on its own?
No. Every real trade is signed by your wallet, and live trading only happens after you deliberately arm and fund it.
Do agents really run when the app is closed?
Yes — paper agents run on Factor's servers 24/7. You get a debrief when you come back.
Is paper trading limited?
No. Paper is free and unlimited, with full history, and doesn't require a subscription.
My agent hasn't traded in days. Is it broken?
Maybe not — but Factor won't make you guess. Open the agent and read why it didn't trade. "No setup" and "below conviction" mean it looked and passed. "Couldn't act" means it wanted to and something in the setup stopped it. And the daily report distinguishes a deliberate sit-out from a day where Factor simply couldn't see.
Does a stronger signal mean a bigger position?
Conviction affects sizing and leverage, but Factor deliberately holds back when it is only partly informed: if some of the signal legs aren't reporting, confidence falls and the agent takes the more conservative route. A half-blind read should not produce a maximum-size trade.
What can the advisor do?
It explains a token's thesis in plain language, grounded in the signals. It never places trades or chooses direction.