Copy trading platforms

What the leaderboard is not showing

Bot	Type	Keys	Price	Copy	Trial	US	Status
ZI Zignaly	Copy trading	pooled	unknown	yes	-	-	-
Superalgos	Self-hosted	self-hosted	free	yes	-	US	-
Cryptohopper	Cloud platform	API keys	$29-129/mo	yes	3 days, no card	US	-
WunderTrading	Cloud platform	API keys	$0-100/mo	yes	7 days, no card	US	-
3Commas	Cloud platform	API keys	$15-140/mo	yes	3 days, no card	US	-
MI Mizar	Cloud platform	API keys	0.1% of volume	yes	-	-	-
Binance Trading Bots	Exchange-native	no keys	free	yes	-	-	-
BI Bitget Copy Trading	Copy trading	no keys	unknown	yes	-	-	-
BY Bybit Copy Trading	Copy trading	no keys	unknown	yes	-	-	-
OKX Trading Bots	Exchange-native	no keys	free	yes	-	-	-
Bitget Trading Bots	Exchange-native	no keys	free	yes	-	-	-
BY Bybit Trading Bots	Exchange-native	no keys	free	yes	-	-	-
GM GMGN	DEX / Telegram sniper	wallet	1% per trade	yes	-	-	-

Live values from our database - details and caveats in each review. "-" = not documented.

Copy trading's pitch is delegation; its reality is a principal-agent problem with a leaderboard. The leader is paid on your wins[3] - so the incentive gradient points toward variance, not durability. The venue shows the survivors[5]. Copy slippage taxes every fill in the wrong direction. None of this makes copy trading bad - it makes the evaluation work real: drawdown history, strategy-change timestamps, sizing consistency, closed-loss publication.

The custody split is the cleaner decision. Exchange-native copy (Bybit[1], Bitget[2], OKX[4]) mirrors trades inside your own account - no third-party API keys exist, so the leak class that cost 3Commas users six figures in 2022[8] is structurally absent. Pooled models (Zignaly's Z-Indexes[3]) place funds with the platform - a different counterparty, evaluated differently. Our tracking adds 6 documented security incidents across this page's operators - custody at major exchanges is not free of risk either, just different risk.

Reading a track record like an auditor

• Returns are the least informative number. Ask: over how many regimes? The bot-vs-hodl backtests show how differently one strategy scores across five BTC regimes - human traders vary more.
• Strategy-change timestamps are the tell. A leaderboard that resets or re-brands after a drawdown is manufacturing survivorship; Bybit and Bitget expose per-trader history[1][2] - read the troughs, not the peaks.
• Profit share + fees set the hurdle. At 10% profit share plus venue fees (0.1%-class per fill[6]), a leader must outperform the market by a real margin before you break even. The bot cost calculator prices the comparison against running your own grid.
• Follower-count rank is popularity, not skill - the same fallacy as stars on GitHub (see our open-source hub for why popularity misleads).
• Set your own stops before following. Inheritance of the leader's risk appetite is not a stop-loss policy.

Which platform for which case

Custody-conscious copying: exchange-native - Bybit's copy market with public trader stats[1], Bitget's larger roster[2], OKX's copy endpoints[4]. No third-party keys, standard venue fees, US access gated.

Performance-fee-only pricing: Zignaly - 10% success fee on net profits, no subscription[3]. Pooled custody; suited to traders who want no exchange account in the loop and accept the platform as counterparty.

Bot-plus-copy hybrids: 3Commas, WunderTrading, Cryptohopper and Mizar run copy/marketplace layers on top of bot automation (our reviews carry the dated plan tables). Flexible, but they reintroduce the API-key chain[8].

The fees under the profit share

The profit share gets the attention; the venue fee schedule does the compounding. Every copied fill pays standard trading fees[6] - on a leader who trades daily, that is the exchange's full schedule applied to your allocation, month after month, before the 10%-class share of wins[3] touches the account. The honest annual bill for copy trading is: venue fees on all fills, plus profit share on winning trades, plus copy slippage on every entry - which is why our bot cost calculator prices copy lanes against running a simple grid on the same capital. For many profiles, a mediocre automated strategy beats a good leader after the fee stack; that comparison is arithmetic, not opinion.

Copy platforms know this, which is why several bundle copy with bot subscriptions and marketplace signal packs - each layer with its own billing. Read the plan matrices in our reviews for the dated numbers before stacking.

What the numbers say

• 13 platforms pass this page's copy filter; 8 of them need no third-party API keys - the exchange-native custody advantage[8].
• 3 offer a trial window and 12 carry a free tier - but copy trading's real trial is a small allocation through one full leader drawdown.
• Median entry price across subscription lanes on this page: ~$22/mo/mo - the fixed-cost layer under profit shares that scale with your success.
• 6 security incidents are documented across this page's operators in our database - exchange custody is the trade you are making for the no-key mirror[8].

A due-diligence routine that fits in an evening

The practical version of "verify the track record", per leader, in order: pull the full return and drawdown chart, not the headline return - Bybit and Bitget surface these per trader[1][2]. Mark any strategy-change or reset events; a record that only shows post-reset performance is unauditable by construction. Check the sizing consistency - a trader who swings from 2% to 50% of book between trades is re-risking behind your allocation. Read the worst week, not the best month: the trough tells you your own drawdown floor if you had followed at size. Finally, sample five closed losing trades and compute what your account would have lost with the venue's fee schedule[6] and the profit share applied - that number, not the headline return, is what you are signing up for.

Then the meta-check: does anything about this leader's marketing (Telegram channels, profit screenshots, urgency) appear in our scam patterns guide? The FTC's patterns[7] cover the copy-trading funnel the same as any other. An evening of this routine filters most of what the leaderboard rewards and the account regrets. Re-run it quarterly: leaders change strategies, venues change fee schedules, and the drawdown you checked last quarter is not the one that matters next quarter. Due diligence in this lane is a subscription, not a purchase.

The bot-vs-leader decision, honestly framed

Both paths delegate strategy to a system you do not fully control - bots to code you configured, leaders to a human you cannot. The comparison axes that actually differ: custody (exchange-native copy removes the key vault; cloud bots require keys[8]), cost (bots bill fixed subscriptions, leaders bill on your wins[3]), and failure mode (a bot fails predictably and visibly; a leader fails quietly by changing behavior). Our grid and DCA hubs price the automated routes with full fee math; this page prices the human route. Between them, the right answer is the one whose failure mode you can monitor - which is a statement about your time, not their quality.

For most readers who arrived here wanting "someone smart to trade for me": the due-diligence routine above is the price of admission either way. A bot you backtest and a leader you audit both demand an evening before they deserve an allocation; only one of them keeps asking for attention afterward.

Warnings

• Survivorship bias is not a flaw the platforms can fix - it is the product. Judge leaders on drawdown depth and regime coverage[5]; assume every leaderboard is curated by attrition.
• The profit share pays variance. A leader who doubles down to recover a drawdown is optimizing their share, not your account. Follow rules you can read, not personalities.
• Copy slippage is real and asymmetric - it worsens losers more than winners. Budget it into the hurdle a leader must clear.
• "Follow the whale" groups on Telegram are the scam extension of this lane - the FTC's patterns[7] and the FCA's bot-branded warnings (see our scam patterns guide) cover the funnel. Real copy venues do not recruit via DMs.
• Custody still matters. Exchange-native copy removes the third-party key vault[8]; the exchange itself remains the counterparty (our /incidents/ timeline tracks both hacks and responses).