EA Builder
A drag-and-drop canvas (plus an AI chat that can do the same work conversationally) for assembling your own MT5 strategy from real indicator and price-action blocks, backtesting it against real historical data, tuning and stress-testing it, and generating a real, ready-to-attach compiled EA -- no coding required.
Canvas & Block Palette
The main workspace: four zones (Long entry, Short entry, Exit long, Exit short) you fill with blocks from a palette of indicators (RSI, MACD, Bollinger Bands, and more), price-action patterns (Fair Value Gap, Order Block, Market Structure, and more), and candlestick patterns, grouped into Trend / Oscillator / Volatility / Volume / Price Action categories.
- Drag a block onto a zone, or use the click-to-add picker above the palette (the only option on touch devices).
- Every block in a zone is automatically ANDed together; add an "Any of" (OR) group when you want alternatives instead of requirements, and nest AND/OR groups inside each other for more complex logic. A NOT-wrapper (negating any condition or group) exists in the underlying data model and the AI chat can produce one, but the canvas editor doesn't yet render or let you visually edit a NOT-wrapped condition after it's added -- treat this as an advanced, chat-only capability rather than a routine workflow.
- Prefer keyboard to dragging? Click a condition's empty operand pill's small round "Assign" button to arm that slot, then click (or Tab and Enter) any palette block to fill it -- the same result as a drag, without a pointer.
- Set a block's timeframe dropdown to something other than "current chart" to compare timeframes in one condition -- e.g. an M15 entry signal filtered by an H1 trend reading.
- Open the full indicator & block encyclopedia below this guide, or take the in-depth EA Builder course, to see every block explained individually.
- A few discrete, event-style blocks (like Fair Value Gap or Order Block) don't support "crosses above/below" comparisons, since they don't have a meaningful previous value to cross -- the condition picker doesn't filter this for you, though: picking a cross operator on one of these blocks anyway gets rejected at generate/backtest time, not blocked in the UI up front.
- Building and editing on the canvas is free and unmetered on every tier, including Free.
Position Modes
A Money & Trades setting that controls how your strategy manages multiple positions: Normal (one position per side, the default), Pyramid (adds to a winning position once it's ahead by a minimum R multiple, up to a capped number of adds), or Grid Martingale (opens a basket that adds to a LOSING position at ATR-spaced intervals with an increasing lot size each level, managed and closed as one whole basket instead of position by position).
- Open the Money & Trades tab and pick a mode from the "Position Mode" dropdown -- this replaced the old "one position at a time" checkbox, so Normal is the equivalent of that checkbox being on.
- Pyramid mode shows its existing minimum-R and max-adds fields, unchanged, just gated behind this dropdown now instead of a separate toggle.
- Grid Martingale mode reveals its own fields: grid spacing (an ATR multiple, never a fixed pip distance), lot size multiplier per level, max grid levels (required -- there's no unbounded mode), basket take-profit/stop-loss/trailing (all in account-currency dollars across the whole basket, not per position), and a fixed lot size fallback used when Risk % is 0 or less.
- In Grid Martingale mode, level 1 opens on your normal entry signal like any strategy; level 2 and beyond only open once price has moved AGAINST the open basket by (grid spacing x ATR), each at a bigger lot than the last -- never on a favorable move, and never past your max grid levels.
- The basket is managed as one unit, not position by position: no per-position stop-loss, take-profit, trailing, breakeven, or aging-out. Live and in a backtest alike, it closes only when the basket's combined dollar profit/loss crosses your basket TP or SL, gives back your configured trailing amount, or your strategy's own Exit Long/Exit Short condition fires.
- Your compiled EA requires at least one of Basket TP, Basket SL, or Basket Trailing set above zero before it will even start -- an exit condition alone does NOT count there, so a config relying only on an exit condition backtests fine but the compiled EA refuses to run. Backtesting is slightly more lenient (it also accepts an exit-condition-only setup, though Basket SL alone still isn't enough on its own there either) -- always set a real Basket TP or Trailing amount before you generate, don't rely on backtest acceptance alone.
- Backtests and the compiled EA count one trade per closed basket (win rate/profit factor are computed on baskets, not individual levels), while the detailed trade log still lists every level so you can see how a basket built up.
- Grid Martingale increases position size while a trade is going against you, so it carries materially more drawdown and margin risk than Normal or Pyramid by design -- keep the lot multiplier and max levels conservative, and always configure a real basket SL or trailing exit.
AI Chat Assistant
A chat that builds and edits your strategy from plain-English descriptions, and can trigger every tool below conversationally instead of you clicking through tabs.
- Describe what you want ("go long when RSI is oversold and price is above the 50 EMA") and it edits the canvas directly.
- Ask "what if" questions (e.g. "what if my stop was tighter") to get a hypothetical backtest without touching your real strategy, or ask it to run any of the Optimize/Analysis tools by name.
- It replies in whatever language you write in.
- Every message costs the same flat chat-message price regardless of what it does; some actions it takes (Optimize, the paid Analysis tools) additionally cost their own compute credits on top, exactly as if you'd clicked them yourself.
- On Free tier it can't load the curated preset gallery even if you ask for one by name -- it explains that and builds you a comparable strategy from the block palette instead.
Discover
A bounded automatic strategy search that builds candidate strategies from scratch -- for when you don't have a starting idea, rather than editing what's already on the canvas.
- Click "Discover a strategy" -- it works on a blank canvas, using your current symbol/timeframe and your existing Money & Trades settings (stop-loss, take-profit, trailing, partials, pyramiding) for every candidate it tries; only the entry conditions are discovered.
- It runs a Thompson Sampling search across the 5 indicator categories (Trend, Oscillator, Volatility, Volume, Price Action), concentrating more of its fixed search budget on whichever category is producing the most promising early results rather than spending it evenly.
- You get the top 5 candidates (each single-direction, long-only or short-only, with at least 10 trades), scored on Sharpe and profit factor, plus which category each came from.
- Click "Load onto canvas" on any candidate to replace your current long/short entry conditions with it -- your exit rules (stop-loss/take-profit/trailing) are left unchanged, with a confirmation step first.
- Costs 6 compute credits -- Discover runs a bounded search of up to 200 candidate backtests (warm-start pulls count against that same 200 cap, not on top of it), capped at 45 seconds of wall-clock time, whichever limit hits first.
Backtesting
A real backtest of your current canvas against historical price data, with a full results breakdown.
- Pick a symbol, timeframe, and optional date range, then click "Run backtest."
- You get trade count, win rate, profit factor, total return, max drawdown, a Sharpe ratio, a Buy & Hold comparison, a breakdown by market regime (trend/chop/high-vol/low-vol), an equity curve, a clickable candlestick trade chart, and a downloadable trade log.
- Results come from historical price data, not your live broker feed -- treat them as an estimate to validate in MT5, not a guarantee.
- Always free and unmetered on every tier.
Optimize
A parameter sweep that tests up to two parameters across many candidate values at once (three search strategies to choose from) and shows you the whole result set, not just the single best combination -- plus robustness checks that tell you whether a result is likely to hold up or was just found by luck.
- Pick up to two parameters (e.g. stop-loss and take-profit distance) and up to 10 candidate values each, then choose a search mode: Grid tests every combination exhaustively; Adaptive narrows in on promising areas across rounds for wide search spaces; CMA-ES learns correlations between the two parameters (e.g. "SL and TP tend to move together") instead of tuning each independently, and usually converges faster once that correlation exists. Adaptive and CMA-ES both force the Pareto scatter view instead of the heatmap, since neither produces a full rectangular grid.
- Optionally test across 2-5 symbols at once ("basket mode", multiplies the cost per symbol) to find settings that hold up broadly rather than one that's secretly tuned to a single symbol, or turn on the walk-forward checkbox to confirm the winner across 4 sequential out-of-sample windows instead of one train/test split -- basket mode and walk-forward are mutually exclusive.
- Grid mode shows a color-coded heatmap -- a broad patch of similar colors means a robust setting, a single bright cell surrounded by weak ones is a warning sign of an overfit spike.
- Adaptive, CMA-ES, and basket runs show a scatter of every candidate scored on return vs. drawdown, with the genuinely best (non-dominated) options highlighted.
- With walk-forward on, you get a Walk-Forward Efficiency badge -- out-of-sample return divided by in-sample return, with a confidence interval: 0.70 or higher reads as robust, 0.50-0.70 as acceptable, below 0.50 as a curve-fit warning. If a candidate's walk-forward windows pass in some market regimes but not others, an "Apply regime gate" button appears that adds a Market Regime Filter condition restricting the strategy to the single regime it most consistently held up in (the button names the specific regime, e.g. "Apply regime gate (trend)").
- Grid sweeps with 4 or more candidates (non-basket) also report a Probability of Backtest Overfitting percentage -- the chance a different historical time-split would have picked a different, worse winner -- alongside a shrinkage-adjusted expected Sharpe ratio that corrects for picking the best of many candidates.
- Clicking any point in the Pareto scatter view applies those parameters to your canvas, with a confirmation step showing exactly what changed (the heatmap view -- the default for Grid mode -- is for visual inspection only; switch to the Pareto view, or use Adaptive/CMA-ES mode which shows it by default, to apply a result with one click).
- Costs 3 compute credits for a single symbol, or 3 credits per symbol in basket mode (6-15 credits for 2-5 symbols); walk-forward and the choice of search mode itself don't add extra cost on top.
Portfolio
A blended backtest across 2-4 of your saved strategies at once, so you can see whether running them together actually reduces risk or just doubles up on the same bet -- plus a closed-form optimal-weight suggestion instead of guessing at weights by hand.
- Pick 2-4 strategies and a relative weight for each (any numbers, e.g. 2/1/1 -- they're normalized to 100% automatically), then run it -- adjusting weights afterward recalculates instantly with no new backtest needed.
- Click "Suggest Kelly weights" for a real closed-form growth-optimal allocation (half-Kelly, computed from each strategy's own return/variance and their correlation) instead of a search -- it fills the weight inputs directly and runs entirely in your browser, no backtest round trip.
- You get each strategy's own individual numbers, the combined blended result, and a plain verdict on whether the blend actually improved on a naive average of the individual strategies (real diversification) or not (they tend to lose money at the same time).
- A correlation matrix shows how related the strategies' returns really are -- anything above 0.9 is flagged as "really one bet twice."
- "Generate all EAs (sized by weight)" compiles a separate .ex5 for every strategy in the blend, each sized to its own weight; "Save this portfolio" keeps the strategy selection and weights for later; "Export report (CSV)" downloads the full numbers.
- Costs 1 compute credit per strategy included in the blended backtest itself (2-4 credits total); Kelly weights, generating the EAs, saving, and exporting are all free on top of that.
Monte Carlo
A simulation that reshuffles your backtest's own real trades into 1,000 new random orderings, to show the range of outcomes that could plausibly happen from trade-order luck alone -- not just the one sequence that happened to occur.
- It appears automatically under any backtest with 10 or more trades -- adjust the position-size multiplier or "ruin" drawdown threshold to see how the range shifts.
- You get 5th/median/95th-percentile ranges for max drawdown and total return, how long a typical recovery from a drawdown takes, and a real probability estimate of hitting your chosen "ruin" threshold.
- Runs entirely in your browser on data you already have -- free and unlimited, no extra request or charge.
Analysis Tools
Ten focused checks you can run on your strategy, each answering one specific question -- from whether individual conditions earn their place to whether your whole edge survives a stricter null-benchmark than a plain backtest.
- Run any of them from the Analysis tab (needs at least one Long/Short condition, except Assess symbol edge which works on a blank canvas) or ask the AI chat to run one by name. Ablate has an extra "Combinatorial (exact Shapley)" checkbox -- slower, but correctly credits/discounts redundancy and synergy between conditions instead of just each one's isolated removal effect.
- Critique red-teams your strategy for regime concentration, fill-order fragility, and whether it's a near-duplicate of an existing curated strategy.
- Mine (good trades) tightens a threshold from your winning trades and validates it out-of-sample; never auto-applies.
- Ablate (conditions) removes each condition one at a time to show which ones are actually earning their place; never auto-applies.
- Stress test costs re-runs your backtest at 1x/2x/3x/5x normal trading costs, and also reports the exact cost-breakeven multiplier -- where profit factor crosses 1.0 -- with a confidence interval, so you know how much execution-cost margin you actually have, not just pass/fail at 5x.
- Check parameter robustness jitters your tuned parameters +/-10% twenty times to tell a genuine edge from a fragile, overfit one.
- Assess symbol edge statistically tests whether a symbol/timeframe has any learnable edge at all, independent of your strategy.
- Suggest a block to add ranks every catalog block not already on your long zone by its Shapley-value contribution, accounting for how it would interact with what's already there, not just added alone -- shows a ranked list, no direct Apply button (add the top suggestion yourself).
- Monte Carlo Permutation Test re-runs your whole strategy (entries AND exits) against synthetic price paths that preserve real volatility clustering -- a stricter test than a plain backtest.
- Exit-logic null-benchmark fixes your real entries and swaps in 3 naive exits (a flat N-bar hold, an opposite-signal exit, an ATR-stop-only exit) with a confidence interval on the P&L difference, to check your exit logic is actually adding value over the simplest alternatives.
- Morris sensitivity screen is an interaction-aware upgrade to Check parameter robustness, covering exit-zone parameters and indicator-choice swaps (e.g. "RSI vs. Stochastic"), not just entry-zone numeric jitter.
- When Stress test, Robustness, or Critique flags a real problem, a "Suggest a fix" button appears, proposing one small concrete parameter change.
- Critique, Mine, Ablate, Stress test, Robustness, Assess symbol edge, Monte Carlo Permutation Test, and Exit-logic null-benchmark each cost 1 compute credit (Critique costs 1 more for every existing curated strategy sharing your symbol, up to 12, it checks against for duplication). Suggest a block to add and Morris sensitivity screen cost 4 compute credits each, reflecting the much larger number of backtests (roughly 130-215 and 90-160 respectively) each one runs to produce a reliable ranking or sensitivity score.
Presets & Curated Gallery
Save/load for your own strategies with full version history, plus (Pro and Pro Max only) a gallery of independently-verified strategies from SmartTrader's own research.
- Save your current canvas as a named preset any time; overwriting an existing one automatically snapshots the prior version first.
- Open a preset's History to diff any past version against your current canvas or restore it, and use Compare to diff any two strategies side by side.
- Loading a curated preset replaces your entire canvas -- save it as your own preset afterward if you want to customize it further.
- Saving, loading, and restoring your own presets is free and unmetered on every tier; the curated gallery itself, and the Compare tool (diffing two strategies side by side), are both Pro/Pro Max subscription perks that a pay-per-use balance can't unlock on Free -- the Compare toggle doesn't even render for Free-tier accounts.
Smart Features
Six opt-in, bounded adaptive behaviors compiled directly into your EA -- automatic risk sizing, trade management, and entry sensitivity that respond to the strategy's own recent performance and detected market regime.
- Open the "Smart Features" tab (between Panel and Presets) to toggle each feature on or off; expand "Advanced" on most of them to tune their numeric bounds.
- Adaptive Risk Governor halves/quarters position size during a rolling drawdown and pauses new entries after a losing streak.
- Regime-Adaptive Management adjusts trailing/breakeven distance and position size based on a detected trend/chop/high-volatility regime -- no tunable parameters, its per-regime multipliers are fixed.
- Dead-Strategy Detector + Shadow Recovery pauses real trading into simulated "shadow" positions once rolling expectancy turns negative, and resumes once shadow performance recovers.
- Bounded Exit-Parameter Adaptation tightens or loosens trailing/breakeven distance based on how much of each winner's peak favorable excursion got given back.
- Reality-Gap Telegram Digest periodically compares live performance to this strategy's own last-run backtest baseline and sends a Telegram digest -- requires Telegram enabled in the Panel tab.
- Bounded Entry-Threshold Adaptation nudges numeric entry thresholds (e.g. an RSI level) tighter or looser based on that side's recent win rate.
- All six default on for a new strategy; every one is bounded and deterministic -- none of them run an unbounded optimizer against your live account.
- Free and unmetered on every tier, same as the rest of the canvas.
Generate & Download
The final step: turning your canvas (or a preset you've customized) into a real Expert Advisor file you can run in MT5.
- Click "Generate" from the canvas at any time -- you don't need to have run a backtest first.
- Optionally enter one or more MT5 account numbers, comma-separated, in "Lock to account number(s)" before generating -- leave it blank to allow any account.
- You get a compiled .ex5 file, not raw .mq5 source -- it compiles server-side, so there's nothing left to build yourself in MetaEditor.
- Drop the .ex5 into your MT5 terminal's Experts folder and attach it to a chart like any other EA.
- An account lock is compiled directly into the file as a fixed constant, not an editable runtime input -- it can't be removed by editing the Expert Advisor's inputs in MT5, and it only lives in the generated file, not in your saved preset or chat session.
- Free and unmetered on every tier, same as backtesting.
Indicator & block encyclopedia
Every indicator and canvas block in EA Builder, explained: what it is, how it works, and how to use it.
Trend
A smoothed line that follows price at a lag, used to read the underlying trend direction without the noise of individual candles.
A moving average with two volatility-based bands above and below it, showing both trend direction and how "stretched" price currently is relative to its recent range.
A series of dots that flip from below price (uptrend) to above price (downtrend), acting as a trailing stop level that accelerates as a trend matures.
A moving average with two bands offset by a fixed percentage rather than by volatility, marking a simple "normal trading range" around the trend line.
A multi-part Japanese trend/momentum system; EA Builder exposes its two most actionable signals as four boolean buffers rather than the full five-line chart.
Oscillator
The Relative Strength Index -- the most common momentum oscillator, bounded 0-100, used to read overbought/oversold conditions.
Moving Average Convergence/Divergence -- a trend-following momentum indicator built from the relationship between two moving averages of different speeds.
A momentum oscillator measuring where the current close sits within its recent high-low range, bounded 0-100 like RSI but calculated differently.
The Commodity Channel Index -- an unbounded oscillator (unlike RSI/Stochastic) measuring how far price has deviated from its statistical average.
The Average Directional Index -- measures trend STRENGTH, not direction; a companion pair of +DI/-DI buffers separately show which direction is dominant.
The simplest momentum measure: how much price has changed over the last `period` bars, expressed as a ratio.
A momentum oscillator nearly identical in concept to Stochastic, but scaled 0 to -100 instead of 0 to 100 (an inherited quirk of its original design).
A momentum oscillator focused specifically on identifying price exhaustion (potential turning points) rather than general trend strength.
The Relative Vigor Index -- measures the conviction behind a trend by comparing where each bar closes relative to where it opened, not just raw price change.
Moving Average of Oscillator -- essentially MACD's histogram exposed as its own standalone indicator (the gap between MACD's main and signal lines).
A momentum indicator comparing a fast and slow simple moving average of the bar's midpoint price, with no configurable parameters.
The Accelerator/Decelerator Oscillator -- measures the rate of CHANGE of momentum itself (Awesome Oscillator's own momentum), catching shifts before AO does.
Measures buying pressure specifically -- how far above a moving average the bar's HIGH reached, isolating bullish strength from bearish strength (its counterpart, Bears Power).
The mirror of Bulls Power -- measures selling pressure by how far below a moving average the bar's LOW reached.
Volatility
Average True Range -- the single most-used volatility measure in EA Builder, used throughout the Money & Trades settings (stop-loss, take-profit, trailing) as the standard unit for sizing distances relative to current market movement.
A statistical measure of how much price is dispersed from its own moving average -- a second, more mathematically literal way (alongside ATR) to read how spread-out recent price action has been.
A direct "is the market compressed right now" reading, exposed as a clean boolean buffer instead of requiring you to eyeball tightening Bollinger Bands yourself.
An open-space filter -- how much room price has to run before hitting its nearest recent swing high/low, measured in volatility units so it's comparable across symbols.
Measures how tightly this symbol is currently moving in lockstep with a reference instrument (DXY, GOLD, or OIL), useful for filtering out trades that are really just a side effect of a stronger correlated market.
A stop-hunt reversal detector -- catches the moment price briefly pokes beyond a recent high/low (sweeping resting stop orders) and then snaps back inside, a classic liquidity-grab pattern.
A session-range breakout detector -- watches the high/low range formed during a specific hours-of-day window (e.g. the Asian session), then flags when price later breaks out of that frozen range.
A market-condition classifier gate -- checks whether the current bar's detected regime (trending, choppy, high-volatility, low-volatility) matches the one you want your strategy to trade in.
Volume
The Money Flow Index -- essentially RSI's overbought/oversold reading but weighted by volume instead of price alone, bounded 0-100.
On-Balance Volume -- a cumulative running total of volume added on up-closes and subtracted on down-closes, used to see whether volume is confirming or diverging from the price trend.
Combines the size of a bar's price move with the volume behind it into a single momentum-with-volume reading -- a big move on heavy volume scores much higher than the same move on thin volume.
A price-scale indicator (like a moving average, not a bounded oscillator) showing where the most trading activity has concentrated recently -- the price levels the market has spent the most volume at, not just the most time.
The Volume-Weighted Average Price -- a running average of price weighted by volume, reset at the start of each trading day, widely used as an intraday fair-value benchmark.
The same VWAP calculation as the standard VWAP block, but anchored to the start of the trading week instead of resetting every day -- a longer-horizon volume-weighted benchmark.
Price Action
A price-action pattern where a 3-candle sequence leaves an untouched gap between candle 1 and candle 3 -- a discrete, event-style signal, not a continuous line like an indicator.
A price-action pattern marking an "impulsive structure break" -- a simplified stand-in for the classic order-block concept (the last opposing candle before a strong move), fired as a discrete event rather than tracking the exact origin candle.
The classic single-candle reversal pattern -- Hammer/Pin Bar (bullish) and Shooting Star (bearish) -- identified by wick-to-body ratio.
The classic 2-candle reversal pattern where the current candle's body fully swallows the prior candle's body.
Classic Floor Pivots -- a price-scale line (the Pivot Point, PP) computed from the previous complete calendar day's high/low/close, a widely-used intraday support/resistance reference.
Break of Structure (BOS, trend continuation) and Change of Character (CHoCH, trend reversal), bundled together as one 4-buffer indicator since both share the same underlying fractal-swing detection and trend-state tracking.
A discrete event marking two consecutive swing highs (or lows) that landed at nearly the same price -- a classic liquidity-pool signature, since equal highs/lows are where stop clusters tend to sit.
A continuous percentage value describing where the current close sits within the most recent confirmed swing range -- below 50% is the "discount" half, above 50% is the "premium" half.
A persistent price zone marking where an impulsive break originated -- price later returning into that zone ("mitigating" it) is treated as a resumption opportunity, the classic order-block-retest concept.
A liquidity sweep followed, within a limited window, by a genuine structural reversal -- distinguishing a real Change of Character from a fakeout sweep that never actually reverses structure.
The Accumulation-Manipulation-Distribution ("Power of Three") session model -- a range forms, gets swept beyond, then price genuinely reverses and moves the other way -- exposed here as its final Distribution stage.
Canvas Mechanics
The canvas's four top-level containers -- Long entry, Short entry, Exit long, Exit short -- each holding the conditions that must ALL be true for that action to trigger.
A `{any: [...]}` group placed inside a zone, where only ONE of its child conditions needs to be true, used to express alternative ways a signal can qualify.
An AND-group (a plain array of conditions) nested inside an OR-group's `any` list, letting you express "(A and B) or (C and D)"-style logic that a flat zone or a single OR-group can't represent.
The four plain numeric comparison operators available on every condition -- `>`, `<`, `>=`, `<=` -- for comparing two operands' values on the current bar.
Two event-style comparison operators -- `cross_above`, `cross_below` -- that fire only on the bar one operand's value crosses the other, instead of staying true the whole time the relationship holds.
A per-operand timeframe pin, available on `ind` (indicator) and `price` operands -- either can be evaluated on a specific timeframe (M1 through MN1) independent of whatever timeframe the chart/backtest itself is running on.
Plain price fields (close, open, high, low, median, typical) and numeric constants as operands for any condition, plus an AI-chat-only formula DSL for custom arithmetic expressions the canvas can't otherwise express.
A wrapper, `{not: <entry>}`, that negates whatever it wraps -- a condition, an OR group, an AND group, or even another NOT -- currently producible only through the AI chat, not a canvas button.
A keyboard-accessible alternative to dragging a palette block onto a condition's operand: arm the slot, then pick a block to fill it.
Settings
The Money & Trades tab's base position-sizing controls -- how much of the account to risk per trade, and the fallback fixed lot size.
The Money & Trades tab's core risk-per-trade controls -- where the stop-loss and take-profit distance, and how (if at all) the stop trails price as a trade moves favorably, are set.
Two related profit-taking controls on the Money & Trades tab -- a single partial-close trigger, and a separate 3-level multi-take-profit ladder -- for scaling out of a position instead of closing it all at once at a single target.
The Money & Trades tab's controls for adding to a winning position ("pyramiding") instead of only ever opening one entry per signal.
A group of Money & Trades tab caps on how many trades the EA can have open or take, and how it identifies its own orders.
Two account-protection filters on the Filters section -- a maximum acceptable spread, and a daily loss circuit-breaker.
Two scheduling filters on the Filters section -- a daily session window, and which days of the week the EA is allowed to trade.
The Filters section's economic-calendar gate, pausing (and optionally closing) trading around high/medium-impact news releases.
A Filters section volatility gate that only allows trading when the current ATR falls within an acceptable range.
Two independent, cosmetic/notification features -- an on-chart status panel, and Telegram trade notifications -- that don't affect trading logic at all.