eatrading.ai v1.0 ← Landing
Reference

Settings Reference — All 41 inputs explained

Every input in the Quant AI Agents EA — type, default value, when to change it, and the watch-outs. Defaults are tuned for XAUUSD on H4 and ship ready to run; you do not need to read this page to use the EA. Open the cards you actually need.

Quick start — most users only touch two fields: paste your API key into 1.2 API key — Analyst (and reuse it across 1.4 / 1.6 / 1.8 / 1.10), then flip 2.3 Auto place orders to true when you're ready. Everything else is already optimized.
Difficulty 41 of 41
No inputs match your filters.

Try clearing the search or switching the difficulty chip back to All.

Group 1

AI Engine — Models, Pipeline, Features

Configure which AI models run each agent role and how the multi-agent pipeline behaves. Most cost-vs-quality tradeoffs live here. The five agents — Analyst, Strategist, Risk Officer, Compliance, and Decision — each take a model and an API key, plus six pipeline knobs at the end.

1.1 Model — Analyst enum GPT-4o-mini

The AI that reads your chart and writes a structured technical report. It picks 3–8 of the most relevant indicators for the current regime, measures them, and emits a JSON observation block consumed by the rest of the pipeline.

When to change

Upgrade to GPT-4o or Claude Sonnet for richer multi-indicator reasoning on choppy ranges; downgrade to GPT-4.1-nano if you re-analyze every 60s and cost is the dominant concern.

1.2 API key — Analyst string (paste)

Provider API key for the Analyst agent. Accepts OpenAI sk-…, Anthropic sk-ant-…, or any OpenAI-compatible gateway key. Stored encrypted in the terminal; never sent anywhere except the configured provider endpoint.

When to change

Paste once and reuse the same key in 1.4, 1.6, 1.8, 1.10 — that is the intended setup for solo users.

Watch out

All 5 agents share the same provider stack; if your key is rate-limited the entire pipeline stalls. Use a key with at least 5,000 RPM when re-analyzing every minute.

1.3 Model — Strategist enum GPT-4o-mini

Synthesizes the Analyst's observations into a directional bias and a candidate entry plan (entry zone, invalidation, R-multiple target). Does not place orders.

When to change

Upgrade only if you've already upgraded 1.1 Analyst — a stronger Strategist on a weak Analyst overfits to noise.

1.4 API key — Strategist string (paste)

Same as 1.2 but for the Strategist. Reuse the same key 99% of the time.

When to change

Use a separate key only when you want per-agent billing dashboards or when running mixed providers (e.g. OpenAI Analyst, Anthropic Strategist).

1.5 Model — Risk Officer enum GPT-4o-mini

Validates the Strategist's plan against position-sizing rules, current open exposure, and structural risk (e.g. proximity to swing highs, asymmetric R:R). Vetoes plans whose effective R:R falls below 1.0.

When to change

The Risk Officer is the agent that most benefits from a stronger model on volatile assets — try GPT-4o on indices or crypto.

1.6 API key — Risk Officer string (paste)

API key for the Risk Officer. Reuse the key from 1.2.

When to change

Provide a dedicated key only if you've enabled 2.11 Verbose log and want clean per-agent traces in your provider dashboard.

1.7 Model — Compliance enum GPT-4o-mini

The macro/news filter. Cross-references the candidate trade window against high-impact events (NFP, FOMC, CPI, geopolitical headlines) and blocks entries inside red-event windows.

When to change

Keep at default unless you specifically want a smaller model — Compliance runs less frequently than the others, so the savings are marginal.

Watch out

Compliance reasoning is only as fresh as the model's training/retrieval cutoff. For breaking news it's a filter, not a guarantee.

1.8 API key — Compliance string (paste)

API key for the Compliance agent. Reuse the key from 1.2.

When to change

Same logic as the other agent keys — separate only for billing clarity.

1.9 Model — Decision enum GPT-4o-mini

The final aggregator. Takes the four upstream agents' outputs, weighs disagreement, and emits the binary decision (OPEN_LONG / OPEN_SHORT / HOLD) plus a confidence score.

When to change

Highest-ROI quality bump on the page. Upgrading only Decision to GPT-4o typically improves trade-selection precision more than upgrading any other agent.

Watch out

The Decision agent is invoked once per re-analysis cycle. On 2.4 Re-analyze every = 60, that is 1,440 calls per day per chart — budget accordingly when upgrading the model.

1.10 API key — Decision string (paste)

API key for the Decision agent. Reuse the key from 1.2.

When to change

If you upgrade 1.9 to a more expensive model, this is the only key whose usage curve will change meaningfully — check it weekly.

1.11 Max indicators per analysis int 5

Hard cap on how many indicator readings the Analyst is allowed to include in a single observation block. Range 3–8. Higher values give the Strategist more context but cost more tokens.

When to change

Bump to 7 for ranging instruments where structure matters; drop to 3 for clean trending markets where one or two indicators dominate.

1.12 Use multi-timeframe context bool true

When true, the Analyst pulls bias from one higher timeframe (HTF = current × 6) before grading the entry timeframe. Adds ~30% tokens per cycle but materially reduces counter-trend trades.

When to change

Turn off only on the highest timeframe you trade (no useful HTF above it) or when explicitly testing pure mean-reversion strategies.

1.13 Use chart vision bool false

When true, the Analyst additionally receives a rendered PNG of the current chart (200 bars, current symbol/TF) and the model is allowed to comment on visual structure (channels, double-tops, S/R clusters).

When to change

Useful with GPT-4o or Claude on visually-driven assets like indices or BTC. Negligible benefit on FX majors.

Watch out

Image inputs cost roughly 3–5× a comparable text call. If you enable this on 1.1 GPT-4o-mini the cost-per-cycle approximately doubles.

1.14 Reasoning depth enum Standard

Controls the system-prompt budget given to each agent. Fast = terse one-shot; Standard = chain-of-thought with self-check; Deep = multi-pass with explicit counter-argument.

When to change

Use Deep only on H4/D1 where you have minutes between cycles. Fast is fine on M5/M15 where speed matters more.

1.15 Cache TTL — seconds int 60

How long to reuse an Analyst observation if neither the price nor the bar index has changed materially. Prevents redundant API calls during sideways minutes.

When to change

Set to 0 when debugging — every cycle becomes a fresh call. Set to 300 on D1 to silence redundant intra-bar calls.

1.16 Telemetry mode enum Compact

Controls what the EA prints to the Experts log. Silent — only errors and trades. Compact — one-line per cycle with the Decision verdict. Verbose — full per-agent reasoning (use sparingly; it floods the log).

When to change

Switch to Verbose when an entry felt wrong and you want to see exactly which agent voted which way.

Group 2

Prompt, Trading & Expert

Strategy selection, the auto-trading toggle, scheduling, and expert-level overrides. This is the group that turns the EA from read-only analyst into something that places live orders — leave 2.3 off until you've watched a few cycles unfold.

2.1 Prompts file string prompts/xauusd_h4.json

Path (relative to MQL5/Files/) to the JSON file containing the per-agent system prompts. Ships with curated presets for the common asset/TF pairs.

When to change

Point at your own file when you want to encode a private strategy library — your prompts override the defaults completely.

Watch out

The schema is strict: a missing role or invalid JSON disables auto-trading on the next reload, not silently. Validate before you ship.

2.2 Strategy preset enum Adaptive

High-level personality applied on top of the prompts file. Adaptive lets the agents pick; Trend, MeanReversion, and Breakout bias the Strategist toward that family.

When to change

Pin to Trend on strongly-trending symbols (e.g. crypto bull runs); pin to MeanReversion for index pairs in tight ranges.

2.3 Auto place orders bool false

The master switch. When false the EA runs the full pipeline and logs decisions but does not place a single trade. When true, every OPEN_LONG / OPEN_SHORT from 1.9 Decision is sent to the broker.

When to change

Flip to true only after running for at least 24h with 2.11 Verbose log on, and confirming the Decision agent is making sensible calls.

Watch out

This setting overrides everything else. A misconfigured stop-loss or lot size will hit your live account the first cycle after you enable it.

2.4 Re-analyze every — seconds int 240

Pipeline cadence. The default of 240s (4 min) suits H4. Match it to your timeframe — 60 for M15, 240 for H1/H4, 1440 for D1.

When to change

Always match to TF: under-sampling misses entries; over-sampling burns API budget without new bar information.

2.5 Trade window — start string 00:00

Earliest server-time hour:minute at which the EA is allowed to open a position. Existing positions are still managed outside the window.

When to change

Set to 08:00 to skip the thin Asian session on FX/Gold; 13:30 aligns with the NYSE open for indices.

2.6 Trade window — end string 23:59

Latest server-time hour:minute at which the EA may open a position.

When to change

Pair with 2.5. A common high-quality window on Gold is 08:0020:00.

2.7 Friday cut-off — hour int 21

On Friday at this server hour the EA stops opening new positions. Set to 0 to disable the cut-off entirely.

When to change

Lower to 18 if you want to flatten before NY-close volume thins. Set to 0 only on instruments with weekend continuation (e.g. crypto).

Watch out

This blocks opens, not closes. Weekend gaps can still hit a position parked over Sunday.

2.8 Magic number int 240501

The integer tag stamped onto every order this instance opens. Used to differentiate multiple EA instances on the same account.

When to change

Always change when you run two or more instances on the same account — pick a unique value per (symbol, timeframe) pair.

Watch out

If two charts share a magic number, the EA cannot tell its own positions apart and basket-exit logic in Group 4 mis-fires.

2.9 Comment tag string QAA

Free-text comment attached to every order. Visible in the broker's order log and in third-party analytics.

When to change

Useful when post-trade tagging matters — e.g. QAA-LIVE vs QAA-DEMO, or per-strategy labels.

2.10 Slippage — points int 30

Maximum allowed slippage on market orders (in points, not pips). 30 ≈ 3 pips on a 5-digit broker.

When to change

Tighten to 10 on liquid majors with low spread; loosen to 50 on Gold during the London-NY overlap.

2.11 Verbose log bool false

Master switch for full per-agent reasoning in the log. Equivalent to setting 1.16 Telemetry mode = Verbose, but easier to toggle from a one-key debug session.

When to change

Turn on briefly when an entry surprised you. Always turn back off — verbose logs are large and rotate the broker's log buffer quickly.

Group 3

Risk Management — Lot + Auto SL (always on)

Position sizing rules and the AI Hybrid stop-loss engine. The Break-Even / Trail safety net runs every 15 seconds locally — regardless of AI status — so a frozen API key cannot leave a position unmanaged.

3.1 Lot mode enum FIXED

How position size is computed. FIXED uses 3.2 as a literal lot value; CUSTOM_RISK sizes the lot so loss-at-stop equals 3.3 percent of equity.

When to change

Stay on FIXED while you're learning the EA. Switch to CUSTOM_RISK only when stop placement is consistent and you trust 3.4/3.5.

3.2 Lot size — fixed double 0.01

Literal lot value used when 3.1 Lot mode = FIXED. The default is the minimum tradable size on most brokers.

When to change

Keep at 0.01 while testing. Even on a $10k account, the realised P&L from the smallest lot is enough to grade the EA's behaviour.

3.3 Risk per trade — % double 1.0

Percentage of current equity risked on each entry. Only used when 3.1 = CUSTOM_RISK. The EA back-solves the lot from this value and the SL distance.

When to change

1.0% is the textbook default. Drop to 0.5% on prop-firm accounts with 5% drawdown caps; never push above 2%.

Watch out

This is risk per single trade. Combined with Group 4's averaging it can compound — verify total open exposure under stress.

3.4 BE / Trail trigger — ×ATR double 1.0

How far (in multiples of current ATR) the price has to travel in your favour before the EA moves the stop to break-even and arms the trailer.

When to change

Lower to 0.7× on fast intraday TFs to lock in profit sooner; raise to 1.5× on D1 to give the trade room.

3.5 Trail distance — ×ATR double 1.5

Once trailing is armed (see 3.4), the SL is dragged this many ATRs behind price. 1.5× is the proven default — wide enough to absorb noise, tight enough to capture swings.

When to change

Tighten to 1.0× when you want to capture quick scalps; widen to 2.0× on swing setups.

3.6 Max concurrent positions int 1

Hard cap on how many simultaneous positions this magic number can hold. Counts toward Group 4 layers as well.

When to change

Stay at 1 while learning. Raise only when you've enabled Group 4 averaging — then the cap should equal 4.2 Max layers.

Group 4

Position Averaging — Layered entry on adverse move

OPTIONAL martingale-family layering. When enabled, the EA strips per-position stops and switches to basket-level exit logic. Risky — leave OFF unless you understand layering and have backtested the exact configuration on your symbol.

4.1 Enable layering bool false

The master switch for averaging. When false, every other input in this group is ignored. When true, the per-position SL is removed at fill and the basket runs to 4.5 Basket TP or 4.6 Kill switch.

When to change

Leave at false until you've explicitly designed and backtested an averaging plan. The defaults below are calibrated, but they are not free money.

Watch out

Enabling this fundamentally changes the EA's risk profile: you replace one bounded-risk trade with a basket whose worst case is set by 4.6, not by ATR.

4.2 Max layers int 3

Maximum number of additional entries the EA will stack when price moves against the basket. 3 means the original position plus up to two adds.

When to change

Stay at 3. Going to 5 doubles the worst-case basket size and rarely improves the average outcome.

Watch out

Each extra layer compounds the basket's loss-at-stop geometrically. Verify on backtest that 4.6 Kill switch is reachable before you exceed broker margin.

4.3 Layer spacing — ×ATR double 1.5

How far (in ATR multiples) price must travel further against the basket before the next layer is added.

When to change

Tighten to 1.0× on stronger trends to lower basket break-even sooner; widen to 2.0× for whippy rangebound assets where you want fewer layers.

4.4 Layer size multiplier double 1.0

Multiplier applied to each successive layer's lot. 1.0 = flat (every layer is the same size). 1.5 = geometric. 2.0 = classic martingale.

When to change

Keep at 1.0. 1.5 is acceptable on a paid backtest; 2.0 is the trade-blow-up dial — never go above 1.5.

Watch out

The basket's worst-case loss scales with the sum of the multipliers, not the count of layers. 1.0 / 1.5 / 2.25 is already a 4.75× single-trade exposure.

4.5 Basket TP — total profit % double 1.0

Cumulative basket profit (as percent of equity at basket open) at which all positions are closed at market.

When to change

Lower to 0.5% for higher hit-rate baskets that close quickly; raise to 2.0% only with a wider 4.3 spacing.

4.6 Basket SL — kill switch % double 10.0

The hard backstop. If the basket's floating loss reaches this percent of equity at basket open, every position is closed and the basket is suspended for the rest of the bar.

When to change

Tighten to 5.0% on prop-firm accounts; 10% is appropriate only for personal accounts where you accept the drawdown.

Watch out

This is the only downside guarantee in averaging mode. Set it to a value you can actually accept — the EA will hit it once.

4.7 Strip per-position SL on layer bool true

When true (default for layered mode), the SL set by Group 3 is removed from each individual position once a second layer fills — only the basket-level exits matter from that point.

When to change

Setting to false creates a hybrid where any single position's SL closes that leg without affecting the basket — useful for asymmetric exit experiments.

Watch out

Hybrid mode is hard to grade — when you're losing money you won't know whether the legs or the basket logic is the cause. Stick to defaults.

4.8 Lock-in BE after N layers int 2

Once the basket has filled this many layers and price recovers above the basket's blended break-even, the basket auto-closes at flat instead of waiting for 4.5 Basket TP.

When to change

Set to 0 to disable break-even rescue and always wait for 4.5. Most users keep at 2 — it converts losing baskets into break-even ones surprisingly often.

Defaults already optimized — try it on demo first

You don't need to read this page to use the EA. Open a demo, paste one API key, and watch a few cycles run before flipping 2.3 Auto place orders on.

Try it on demo — defaults already optimized ← Back to overview