IT Realms
IT Realms Case Study

ChameleFX — Algorithmic Trading Intelligence

An enterprise-grade AI trading platform engineered to execute, adapt, and protect — autonomously operating across 8 strategies on MetaTrader 5 with multi-layered risk management and real-time AI signal validation.

FinTech AI-Powered MetaTrader 5 11-Layer Risk Gates
8Trading Strategies
11Risk Gate Layers
AIGemini Signal Scoring
MT5Live Execution Engine
THE PROJECT

A Trading Engine That Thinks Before It Trades

ChameleFX is an enterprise-grade algorithmic trading system built natively for MetaTrader 5 (MT5) — the industry-standard platform used by institutional traders and proprietary trading firms worldwide. The system is engineered not just to execute trades automatically, but to reason about market conditions before entering any position.

At its core, ChameleFX runs 8 distinct trading strategies simultaneously across multiple timeframes — each strategy independently confirming signals before passing execution authority to the order management engine. Every potential trade is then scored by a Google Gemini AI model, which evaluates the macro market context and applies a confidence multiplier to the position sizing calculation.

Before any trade reaches the broker, it must pass through an 11-layer risk gate — a sequential approval chain covering drawdown limits, volatility filters, session restrictions, correlation exposure, Kelly Criterion position sizing, and more. The result is a system that trades aggressively when conditions are ideal, and goes quiet when they are not.

ChameleFX Algorithmic Trading Dashboard — Strategy Overview
CORE STRATEGIES

8 Strategies. One Unified Engine.

Each strategy operates independently with its own entry logic, multi-timeframe confirmation, and exit rules. All 8 feed signals into the central risk and execution pipeline.

01 — STRATEGY
Trend Following

Identifies and rides sustained directional moves using EMA crossovers, ADX trend strength confirmation, and higher-timeframe bias filtering. Enters on pullbacks within the dominant trend.

02 — STRATEGY
Mean Reversion

Exploits overextended price deviations from statistical norms using Bollinger Band extremes and RSI divergence. Targets the return to mean with tight stop management.

03 — STRATEGY
Momentum Breakout

Captures acceleration events — breakouts from consolidation zones and key structure levels — confirmed by volume surge and momentum oscillator agreement across timeframes.

04 — STRATEGY
Scalping Engine

High-frequency micro-structure strategy operating on M1/M5 timeframes. Exploits bid-ask spread dynamics and order flow imbalances with precision entries and fixed pip targets.

05 — STRATEGY
News Event Trader

Monitors economic calendar events and adjusts position exposure pre/post high-impact releases. Trades directional moves from sentiment shifts after major data prints.

06 — STRATEGY
Grid Recovery

Structured grid system with dynamic level spacing that adapts to volatility. Includes compulsory drawdown ceiling per grid instance — the entire grid closes if the ceiling is breached.

07 — STRATEGY
Support & Resistance

Algorithmically identifies key price structure levels from historical swing data. Enters on reaction signals at these zones with multi-candlestick pattern confirmation.

08 — STRATEGY
AI Signal Strategy

A dedicated strategy driven purely by Google Gemini AI analysis. The model processes current market data, macro indicators, and sentiment to generate directional bias scores for execution.

RISK ARCHITECTURE

11 Gates. Zero Compromises.

Every trade passes through a sequential 11-layer approval chain before execution. If any layer rejects the signal, the trade is silently discarded — never overridden.

Maximum Drawdown Limit
Account-level daily and session drawdown ceiling. Breaching this threshold suspends all trading until the next session reset.
Volatility Filter (ATR Gate)
Uses Average True Range to measure current market volatility. Rejects trades when volatility is too low (no edge) or too high (excessive spread and slippage risk).
Session Time Filter
Restricts trading to defined market sessions per instrument — London open, New York session, Asian overlap windows — avoiding low-liquidity dead zones.
Correlation Exposure Control
Monitors open positions across correlated pairs. Prevents stacking directional exposure — e.g., simultaneous long EURUSD and long GBPUSD during dollar weakness.
Kelly Criterion Sizing
Position size calculated using the Kelly Criterion formula applied to each strategy's historical win rate and reward:risk ratio — scaling size to edge, not emotion.
Max Open Positions Cap
Hard limit on concurrent open trades, both per strategy and system-wide. Prevents over-leveraging the account when multiple strategies align simultaneously.
Economic Calendar Block
Queries live economic calendar data and suspends new trade entries within a configurable window before and after high-impact news events.
Spread Filter
Rejects entries when live broker spread exceeds an instrument-specific maximum. Prevents execution at unfavourable cost during low-liquidity spikes.
Profit Target Lock
Optional daily profit target that, once reached, suspends new entries for the remainder of the session — locking in gains and preventing overtrading.
Equity Protection Trailing
Account-level trailing stop on total equity. As total equity grows, the floor rises — preserving profits on a winning day even if the final trades turn against the system.
AI Confidence Threshold
Google Gemini must return a confidence score above the configured threshold for the trade to pass. Low-confidence AI assessments block execution regardless of technical signal quality.
AI INTELLIGENCE

Powered by Google Gemini

Every trade signal is reviewed by a live AI model before execution. ChameleFX treats AI as a validation layer — not a black box — with full control over how much weight the model carries.

Real-Time Signal Scoring

When a technical strategy fires a signal, ChameleFX packages the current OHLCV data, indicator readings, timeframe context, and open position summary into a structured prompt. This is sent to the Google Gemini API, which returns a directional bias score (bullish / bearish / neutral) and a confidence percentage. Only signals that match the intended direction AND exceed the confidence threshold proceed to execution.

Confidence-Weighted Position Sizing

The Gemini confidence score is used as a position size multiplier — applied on top of the Kelly Criterion base size. A high-confidence AI reading on a strong technical signal increases position size proportionally. A lukewarm AI signal reduces it. This means the system automatically bets more when both technical and AI indicators agree, and less when they diverge.

Market Sentiment Analysis

Beyond raw price data, Gemini is prompted with market sentiment context — recent economic events, major currency themes, and geopolitical factors relevant to the instrument being traded. This gives the AI layer a macro-awareness that pure technical analysis cannot provide, allowing ChameleFX to avoid technically valid setups that conflict with dominant fundamental forces.

Telegram Real-Time Alerts

Every trade event — entry, modification, partial close, full close, and risk gate rejections — triggers a structured Telegram notification. The alert includes strategy name, pair, direction, entry price, SL/TP levels, position size, AI confidence score, and account equity status. Complete visibility without being tethered to a terminal screen.

ChameleFX Live Trade History — MetaTrader 5 Terminal
LIVE EXECUTION

Every Trade Logged. Reviewed. Optimised.

ChameleFX maintains a full execution log inside MetaTrader 5 — every entry, exit, partial close, and modification is recorded with timestamp, strategy tag, and AI confidence score at the time of execution.

This log feeds directly into the strategy performance dashboard, allowing each of the 8 strategies to be individually evaluated — win rate, average RR achieved, slippage incurred, and drawdown profile per session type. Underperforming strategies are identified quickly and can be paused or recalibrated without affecting the others.

  • Per-strategy win rate and expectancy tracking
  • AI confidence score logged against outcome per trade
  • Session-type breakdown (London, NY, Asian, overlap)
  • Real-time Telegram notification for every execution event
TECHNOLOGY

The Stack Behind ChameleFX

Python 3.12 MetaTrader 5 API Google Gemini AI PyTorch FastAPI Telegram Bot API PostgreSQL Pandas / NumPy Economic Calendar API Kelly Criterion Engine VPS Deployment Git CI/CD
BUILD YOUR AI SYSTEM

Need a Custom AI Trading System?

Whether you need an algorithmic trading engine, an AI signal platform, a backtesting framework, or a complete quantitative research system — we have built it and we know how to build it right.

Chat on WhatsApp