How to Copy Trades Across Platforms: Tradovate, NinjaTrader, Rithmic & ProjectX (2026)

You passed three evals at three firms and now your funded capital is scattered across three stacks that do not talk to each other. One firm handed you a Rithmic connection you reach through NinjaTrader. Another runs on Tradovate. The third locked you into TopstepX. You place one trade and you want all three to mirror it. Most copiers cannot do that, because most copiers live inside a single platform and can only see accounts that platform is already logged into.

That is the whole problem in one sentence: a copier sees what its host can reach, and no single front-end reaches every gateway. This guide is the practical version of solving it. Pair by pair, with the latency math and the symbol traps that quietly turn a "successful" copy into the wrong size.

Why your accounts end up on different platforms in the first place

Nobody chooses to spread their book across four stacks. The prop firms choose for you. Each firm strikes its own connectivity deal, and the platform you get is whatever that firm licenses. Apex hands many traders a Rithmic connection. A different firm routes through Tradovate. Topstep moved everyone to TopstepX. You end up holding accounts on incompatible plumbing not because it is optimal but because that is what each rulebook permits.

The key thing to understand is that these names are not all the same kind of thing. Rithmic is a low-latency connectivity gateway, not a charting app. You never trade "in Rithmic" the way you trade in a chart window. You connect a front-end to it: NinjaTrader, Sierra Chart, Bookmap, Quantower, R|Trader Pro, ATAS, Tradesea. Tradovate is both a gateway and a web front-end, so you can trade in Tradovate directly or route a Tradovate account through NinjaTrader or TradingView. NinjaTrader and Quantower are front-ends that sit on top of Rithmic or Tradovate; they are not gateways themselves.

So when someone says "I want to copy from NinjaTrader to Tradovate," the real question is which gateway each leg sits on, not which logo is on the window.

Native group trading vs a true cross-platform copier

There are two completely different ways to mirror orders, and conflating them is the single most expensive mistake here.

The first is platform-internal copying. NinjaTrader has group and sync trading. Replikanto is a NinjaScript add-on that lives inside NinjaTrader 8 and replicates orders between NinjaTrader 8 accounts. These are fast and cheap, but they share one hard limit: they can only move orders between accounts NinjaTrader itself is logged into. Replikanto, being a NinjaScript add-on, needs NinjaTrader 8 plus its own license running on every leader and every follower machine.

The second is an API-level or cloud copier. It does not live inside a trading terminal. It connects to each gateway's API independently, reads fills on the source, and writes equivalent orders into destinations on entirely different stacks. That is the only class of tool that can bridge a Rithmic firm to a Tradovate firm to a TopstepX firm in one group.

The platform pairs traders actually need

In practice, four pairings cover most real setups. Their difficulty is not equal, and the marketing rarely tells you which ones are hard.

The single-Rithmic-connection limit is the ceiling people slam into, and it is a NinjaTrader constraint, not a Rithmic one. NinjaTrader allows only one Rithmic login per instance but many Tradovate logins simultaneously. So "copy between two Apex Rithmic accounts in one NinjaTrader" does not work as a checkbox. You need two VPS instances, or you put one leg on Tradovate, or you use an external copier. Anyone who tells you Rithmic-to-Rithmic local copy is a simple toggle has not tried it with two firms.

Tradovate to NinjaTrader, step by step

This is the friendly pair. NinjaTrader talks to Tradovate over a direct API integration, and it tolerates multiple Tradovate connections in one instance. If both your source and destination accounts route through Tradovate, you can sometimes do this inside NinjaTrader alone. If the destination is a NinjaTrader Brokerage account or a Rithmic account, you cross a gateway boundary and need an API copier.

1. Decide which account is the leader. The leader is wherever you place the original order.
2. Confirm the destination's real gateway. A "NinjaTrader" follower might actually be a Tradovate account shown in NinjaTrader, or a separate Rithmic or NinjaTrader Brokerage feed. The gateway, not the window, determines whether one platform can bridge it.
3. If both legs are Tradovate, add both Tradovate connections to one NinjaTrader instance and use its sync trading, or point an API copier at both.
4. If the legs differ, register both accounts with an API copier that holds independent sessions to each gateway. Tradovate and NinjaTrader are both first-class targets for a copier built this way.
5. Set the symbol mapping and the per-follower size, then run one micro test trade before anything real.

If you also route through charts, note that TradingView can send live futures orders to several brokers via paid add-ons. The NinjaTrader-to-TradingView order-routing add-on is listed around $9.99/month with Level I CME data, and Tradovate integrates with TradingView directly. That means a TradingView chart can act as a leg in the same group, which is the practical answer to copying NinjaTrader to TradingView futures.

Rithmic to ProjectX (and back)

Here is where 2026 broke a lot of old advice. ProjectX moved to an exclusive licensing arrangement with Topstep early in 2026, and every other firm that had relied on it lost access soon after. As things stand there is no standalone ProjectX download and no other firm licenses it, so the ProjectX experience today effectively means TopstepX. Licensing deals shift, so confirm the current state before you architect around it.

So "Rithmic to ProjectX" in 2026 really means bridging a Rithmic prop account into a TopstepX account at Topstep. No single front-end reaches both the Rithmic gateway and the TopstepX gateway, which rules out the in-platform copiers entirely. You need an API-level copier that connects to each gateway separately and normalizes the order between them.

Treat any Topstep account as its own island. If a 2026 article lists "ProjectX" as a general copier target alongside Rithmic and Tradovate, it was written before the walls went up.

The reverse direction, TopstepX as the leader copying out to Rithmic followers, has the same requirement: the copier has to hold an independent TopstepX session and an independent Rithmic session at once. Before you build any of this, confirm Topstep currently permits copier use on your account tier. The deeper compliance picture lives in our piece on whether you can copy trade multiple prop firm accounts, and the gateway comparison is in Rithmic vs ProjectX data feeds. For the platform-specific surfaces, see Rithmic, ProjectX, and TopstepX copier pages.

Why latency, not features, decides cross-platform fills

Lowest latency and widest reach are opposing goals, and the marketing numbers hide it. An in-process local copier hits about 1.6ms (Copilink advertises that) because the order never leaves the machine. But it inherits every connectivity limit of its host platform. A cross-stack cloud copier publishes higher figures, roughly 17ms to 50ms, precisely because it bridges separate gateways and stays up 24/7 without your PC. Tradesyncer cites 30 to 50ms, Tradecopia around 40ms, and Thor states 17ms server-side. Those are vendor claims, not audited benchmarks.

Walk a real latency budget. A scalper's edge on a fast NQ break might be worth a single tick, and on the micro that is $0.50, on the full contract $5. If the copy leg adds 40ms versus 17ms, the question is not "is 23ms a lot" in the abstract, it is "does the book move more than a tick in 23ms during a news spike." Usually no. Occasionally, at 8:30am on CPI, yes. That is the honest tradeoff: the cloud copier you can reach all three firms with is slightly slower than the local copier that can only reach one.

If your whole book happens to sit behind one reachable feed and you are colocated near the exchange, a local copier can genuinely beat a server-side one on raw speed. Thor does not pretend otherwise. A Chicago-colocated local setup on a single stack will win the millisecond race. The moment a second, unreachable stack enters the picture, that advantage evaporates because the local copier simply cannot place the order at all.

Symbol mapping and contract differences across platforms

This is where "it copied but the wrong thing happened" bugs live. Futures symbols are built from a root, a month code, and a year, sometimes with a slash or an exchange suffix. The CME month codes run F=Jan, G=Feb, H=Mar, J=Apr, K=May, M=Jun, N=Jul, Q=Aug, U=Sep, V=Oct, X=Nov, Z=Dec, skipping I, L and O so they are never confused with numerals.

A cross-stack copier must normalize that string or the destination leg never fills. Worse, it must also map the contract multiplier. ES and MES are a 10x notional difference. Good copiers handle full-to-micro mapping (ES to MES, NQ to MNQ) and auto-size quantity, which lets a trade on a mini chart mirror as micros in a smaller follower. A mis-set ES-to-MES mapping or a stale roll month produces fills that look successful but are one tenth or ten times the intended size. That is the quiet killer, because nothing errors out.

Per-follower scaling so a 50K and a 150K account both stay compliant

Here is the worked example that ties size, mapping, and compliance together. Say you hold a leader Apex Performance Account on a Rithmic stack reached through NinjaTrader, and you want to mirror to 9 follower PAs. Your leader's platform shows the June 2026 E-mini S&P as ESM6 (root ES, month code M for June, year 6 for 2026). You go long 1 ES at 5,300.00 and the copier fans 1 contract to each of 9 followers, so the firm-wide position is 10 ES.

ES is $50 per point. If the market moves +4.00 points, from 5,300.00 to 5,304.00, per-account P&L is 4.00 x $50 = $200, and across all 10 accounts that is 10 x $200 = $2,000 gross. Clean on paper. Two things bend it in reality. First, when 10 market orders hit a liquid book at the same instant, the later fills can slip a tick or more, so realized aggregate P&L usually lands a touch under $2,000. Second, if one follower sits on a Tradovate stack that shows the contract as ESM26 or /ES, the copier must normalize the symbol or that leg never fills.

Now the compliance twist. Suppose for risk control you map ES to MES (micro, $5 per point) on the smaller 50K follower. That leg makes 4.00 x $5 = $20, not $200. The aggregate number is no longer 10 identical legs, so verify the multiplier mapping before trusting any firm-wide total. Per-follower scaling is exactly how you keep a 50K and a 150K account inside their own drawdown rules from one master signal: different multipliers, different caps, different symbol mappings, all per account.

The rules bound this. Apex permits copy trading between your own Performance Accounts: one leader PA plus up to 19 followers, with a 20-account total cap, and you may not copy another person's account or sell signals. Topstep allows copier use up to $750K buying power and up to 5 active funded accounts, but prohibits hedging across accounts and coordinated trading with others. Copying your own funded accounts is broadly fine; signal-selling to accounts you do not own is not. These parameters change often, so check the current rulebook with each firm.

Paper-test the bridge before you go live

Every cross-stack bug shows up cheaply in sim and expensively in production. Run the bridge end to end on simulation accounts first, with one micro contract, and watch the destination leg actually fill at the size you expect.

The checklist you can act on today:

• Identify the real gateway behind every account, not the window logo.
• If two legs are both Rithmic at different firms, plan for two instances or an external copier from the start.
• Confirm each firm currently allows copier use on your tier.
• Set per-follower multiplier, max position, daily-loss kill switch, and symbol mapping before the first live trade.
• Fire one MES or MNQ test trade and reconcile the fill, size, and direction on every leg.
• Only then scale to full contracts and more followers.

This is the angle where a server-side copier earns its keep. Thor copies across 11+ platforms from one master, so a single fill can fan out to followers on Tradovate, NinjaTrader, Rithmic, and TopstepX at once, in roughly 17ms as advertised, without your PC staying on. It is the API-level class of copier, which is the only class that can see accounts on different gateways. If your entire book sits on one reachable feed and you live next to the exchange, a colocated local copier may still beat it on raw speed. The cross-platform futures copier question is about reach first. For deeper comparisons, see our roundup of the best futures trade copiers and the breakdown of MT4 vs MT5 vs cTrader for the CFD side, or check the FAQ.

Frequently asked questions