Best Prop Firms for Running an ICT MT5 EA (2026 Ranked List)

Quick Answer: The Best Prop Firms for an ICT MT5 EA Right Now

FTMO, on the Swing account type, is the default pick for most ICT EAs. Its maximum-loss floor is set from the starting balance instead of trailing behind new equity highs, the Swing type carries no news-release restriction and lets positions run overnight and over the weekend, and its automation limit is published as a number you can count your own EA against. No pre-approval, no source code to hand over.

FundedNext is the better first stop if your EA habitually sits through large unrealised excursions, because its evaluation programs are the ones most worth checking for a daily limit read off closed balance rather than live equity. It also bans very little strategy logic — with two exceptions that catch ICT builders off guard, covered below.

Two conditional picks round it out. The5ers is excellent if you wrote the EA yourself, and unusable if you did not: it requires you to own the source code. E8 Markets is the pick when you want the automation ceiling written down in plain numbers before you spend anything.

If you remember one line from this page, make it this one: choose on how the daily loss is measured and whether the maximum-loss floor trails. Everything else is a preference.

How We Ranked Them: Four Criteria an ICT EA Actually Feels

This ranking is built from each firm's own published rule text, read against the behaviour of an ICT-style EA — not from a leaderboard, a rating score, or a firm's marketing badge. Where a rule is program-specific, we say so, because these firms change parameters between programs more often than they change brands.

1. How the daily loss is measured

An ICT entry sits at a level, and price frequently trades through that level before honouring it. The position spends time deeply underwater and then closes green. A firm that checks the daily limit against live equity counts the worst point of that excursion the moment it happens. A firm that checks it against the day's closing balance never sees it at all.

Same trade, same EA, two different verdicts. This is the single biggest reason a strategy that passed on one firm breaches on another, and it is why the daily-limit basis outranks every headline number in this comparison. Work out how much unrealised loss your own worst historical trade would have shown at its deepest point, then check that figure against the limit you would be trading under — our prop-firm drawdown calculator does the arithmetic. How you then hold the EA under that line is a separate discipline of its own.

2. Whether the maximum-loss floor trails

The second rule is the maximum-loss floor. A static floor is fixed at the starting balance and never moves. A trailing floor rises with every new equity high, so your accumulated buffer shrinks the moment you become profitable.

Structure-based stops make trailing floors particularly hostile. Your stop distance is set by where the swing low actually sits, not by a fixed pip count, so your per-trade drawdown varies from trade to trade. A trailing floor asks a variable-risk system to fit inside a shrinking envelope — a bad pairing that has nothing to do with the quality of the edge.

3. What the automation ceiling actually counts

Every serious firm limits how hard an EA may hammer the trade server, and most express it as a maximum number of server requests per day. Read that carefully, because it counts requests, not trades. An ICT EA that maintains pending limit orders at fair-value gaps and re-prices them on each structure update can generate hundreds of modifications in a session while opening three positions.

That is the trap: your EA is not high-frequency by any trading definition, but it can look hyperactive to a rule that counts messages. The firms worth shortlisting are the ones that publish the ceiling as a number — vague "no excessive activity" wording gives you nothing to engineer against. The related permission to check is algo trading permission on the account type itself, since some programs allow automation only on specific platforms.

4. Whether the news blackout intersects your killzones

ICT entries cluster around session opens, which is exactly where scheduled releases land. A blackout window of a couple of minutes either side of a high-impact release is common, and it is not a compliance detail for an unattended EA — it is a functional constraint that can silence your best window. Either the firm offers an account type without the restriction, or your EA needs a news filter and a trading session filter tight enough to sit out those minutes cleanly.

One thing that is easy to miss: your killzone hours are defined in GMT, but your EA's hour inputs run on broker server time, which is usually GMT+2 or GMT+3 and shifts with daylight saving. Move the same EA from one firm to another and its windows move with it unless you re-point them — the same class of problem as clock drift between systems, with the same silent failure mode.

And then payout terms

Payout terms rank fourth for a reason: they only matter after the first three let you survive. When you do compare them, compare the whole shape — the split, how it scales, the payout cycle, and any cap on the first withdrawal — rather than the headline percentage, and run your own numbers through a payout split calculator instead of trusting a comparison graphic.

Deliberately excluded from the ranking: challenge price (a cheap challenge under a hostile drawdown model is not cheap), the headline profit split alone, star ratings, and any firm's own "EA-friendly" badge. A yes/no answer on whether EAs are allowed is also not a ranking criterion — it is a prerequisite, and it is surveyed in depth elsewhere.

The Best Prop Firms for Running an ICT MT5 EA, Ranked

FTMO — best for killzone EAs that must act around news

FTMO's position at the top rests on three things an ICT EA specifically wants. Its maximum loss is measured from the starting balance rather than trailing your equity. Its Swing account type has no restriction on trading during news releases and no requirement to be flat overnight or over the weekend, which removes the single most awkward constraint for an EA built around session opens. And its rule text states the automation ceiling as a daily request count, so you can instrument your EA and know where you stand instead of guessing.

There is no approval process and no source code to submit; the firm's own position is that your style is your business as long as the trading is legitimate and does not resemble a forbidden practice.

Two caveats. The daily limit is the equity-based kind, so criterion one cuts against FTMO — an EA that habitually sits through deep excursions will feel that. And FTMO caps how much capital can sit behind the same strategy across all your accounts, which matters if you bought a widely sold third-party ICT EA: other people are running the identical logic, and the firm treats that as one strategy.

FundedNext — best for EAs that sit through deep excursions

FundedNext is the one to check first when your EA's problem is unrealised loss rather than realised loss. It runs several distinct programs, and they do not share a drawdown model — some evaluation programs use a static maximum loss while an express-style program uses a trailing one — so read the program page for the exact plan you intend to buy rather than assuming the brand behaves one way.

It is also unusually permissive on strategy logic, and it is the only firm here that pays you a share of the profit you generate during the evaluation itself, which meaningfully reduces the real cost of a challenge you pass.

The caveats are specific and catch ICT builders regularly. Automation is allowed on MT4 and MT5 but not on its cTrader or Match-Trader offerings. EA access is a paid add-on rather than a default. EAs marketed as challenge-passing tools are prohibited outright, as are EAs wired to Telegram or WhatsApp — so if your ICT EA pushes signal notifications to a Telegram channel, strip that module before you deploy. And, as at FTMO, one strategy carries a capital allocation cap across your accounts.

The5ers — best for an EA you wrote and control yourself

The5ers is the strictest firm here about what an EA is, and that strictness is a feature if you are a builder. Its rules require the trader to own the EA's source code, and they require stop-losses to be visible in the platform rather than held in "stealth mode" by the EA. If you wrote your ICT EA, both are trivially satisfied and you get a firm whose programs include genuinely permissive variants — some carry no minimum trading days and no news restriction at all, which suits a low-frequency session EA that may sit out a whole week.

The banned list is broad and worth reading in full: tick scalping, latency and hedge arbitrage, copying another person's signals, emulators, and any EA that overloads the trade server with excessive requests. A pure ICT structure EA is nowhere near those categories, with one exception — the server-request rule, which brings you straight back to criterion three if your EA re-prices pending orders aggressively.

The disqualifier is simple. If you licensed a third-party ICT EA and do not hold the source, The5ers is not a candidate. That is worth knowing before you pay, not after.

E8 Markets — best for knowing the ceiling before you buy

E8's appeal for an automated trader is that the limits are stated as numbers rather than adjectives: a daily ceiling on server requests, a ceiling on positions per day, and an expectation that a share of your trades are held longer than a minute. For an EA developer, a published number is worth more than a friendly policy, because you can test against it.

E8 also requires that your strategy be unique to you rather than shared across many accounts, and it prohibits high-frequency and arbitrage approaches outright. The news window is on the tight side, so an EA that trades releases will need the filter discipline described above. Read the payout terms closely too — the fine print around caps on withdrawals is the part reviewers most often skip.

Which Firm Fits Your ICT EA's Trading Style?

The ranking above is the general case. Your EA is a specific case, and two questions about its behaviour will resolve it faster than any feature list.

Underneath the branches sit two very different automated profiles that get lumped together as "ICT bots".

The first is the session-only structure EA: it wakes for one or two killzones, needs a sweep plus a displacement plus a return into a zone, and takes fewer than five trades a week. Its risk is not frequency at all — it is the depth of the excursion it tolerates and the pending-order churn it generates while waiting. Rank firms for it on daily-limit measurement first, request ceiling second.

The second is the higher-frequency SMC scalper: it takes internal-structure entries on lower timeframes, holds for minutes, and trades most sessions. This profile lives much closer to the lines every firm draws around minimum hold times, request counts and high-frequency behaviour, and it is far more sensitive to execution speed and to slippage at the session open. For that profile, platform and execution quality outrank the drawdown model.

The row that surprises people is the first one. Your EA can be the calmest system on the server by trade count and still sit at a third of the request ceiling, purely from re-pricing orders it never fills. Log your own numbers over a normal week before you decide — and while you are at it, check that your lot sizing still fits the account's cap with a maximum lot size calculator, because structure-based stops produce variable position sizes.

What to Verify Before You Buy a Challenge

All of that starts with confirming a firm's exact EA policy in its own current rule text, not in a review — including this one. Prop firms revise programs frequently, and a rule that changed last month invalidates every ranking published before it.

Three items on that list deserve a sentence each. The netting versus hedging question is a genuine breaker: if your EA scales into a position with multiple orders, a netting account will merge them into one position and your partial-exit logic will not behave as tested. Symbol suffixes are the second: many firms serve EURUSD.r or similar, and an EA with hardcoded symbol strings simply does nothing until you fix its symbol mapping. Third, run the EA on the firm's own demo server for a full week — a forward test on live infrastructure catches server-time offsets, symbol names and request throttling that the strategy tester never will, and it costs nothing.

Two questions this page deliberately does not answer: whether a given firm permits EAs at all, in rule-text detail, and how FTMO's specific compliance requirements apply to an ICT system. Both are surveyed separately and in more depth than a ranking should carry. The same goes for the mechanics of holding an automated system under a daily loss line, or through a news blackout — each is its own subject.

So is keeping it inside a consistency rule, which decides how evenly the profit it makes is allowed to arrive.

Finally, the money question. A challenge fee is a real cost with an uncertain outcome, and no ranking changes that; price the attempt honestly with a challenge cost calculator before you commit, and read our risk warning if you are weighing this as an investment rather than as a test of a system you have already validated. Historical or backtested results describe the past, not the account you are about to buy.

Final Verdict

For most traders arriving here with a working ICT EA in hand, this was never really a choice between brands. It is a choice between measuring instruments: read how the daily limit is taken and whether the maximum-loss floor trails, hold those two answers against how deep your own EA's excursions run and how many requests it sends while it waits, and the firm falls out of the arithmetic rather than out of a preference. The split, the scaling path and the price of the attempt only begin to matter once those two rules have let the EA survive its first bad week.

What no firm on this list will do is compensate for a system that has not been forward-tested on its own infrastructure. Choosing the firm is the smallest part of passing a prop firm challenge with an ICT EA — the rest is the risk model running underneath it and the discipline to leave the EA alone once it starts. Keep your VPS for the EA close to the firm's server, point your session windows at the right offset, and let the strategy do what you already know it does.

FAQ

Will my killzone times still be right at a new prop firm?

Usually not without a change. Killzone windows are defined in GMT, while your EA's hour inputs read the broker's server clock, which typically runs at GMT+2 or GMT+3 and shifts with daylight saving. Two firms with different server offsets will fire the same EA at different times of day. Re-point the hours after every move and confirm the first session on a demo account.

Is a third-party ICT EA riskier at a prop firm than one I built?

Yes, in two concrete ways. Firms cap how much capital can sit behind a single strategy across all accounts, and a widely sold EA means many traders running identical logic — which the firm can treat as one strategy. Separately, at least one firm requires you to own the EA's source code outright, which a licence does not give you. Both are answerable before you buy.

My EA barely trades. Can it still breach an automation limit?

It can. The common ceilings count server requests, not trades. An ICT EA that keeps pending orders at fair-value gaps and re-prices them as structure updates can send hundreds of modifications during a session while opening two or three positions. Log the request volume of a normal week and compare it to the published ceiling before you assume you are safe.

What usually breaks first when an EA moves between firms?

In order: symbol names, because suffixes differ and hardcoded strings silently match nothing; the server time offset, which moves every session window; and the account's netting or hedging model, which changes how multiple orders on one symbol behave. All three surface within a day on the firm's demo server and none of them surface in the strategy tester.

Sources & Further Reading

Want to go deeper? These independent, authoritative sources shaped this guide — each one is worth reading in full:

• FTMO FAQ - which instruments and strategies are allowed, including EAs
• The5ers - Prohibited Trading Practices
• FundedNext Help Center - Is EA allowed in FundedNext