Complete user guide for the MT5 Liqbot AI v10 — your professional-grade, AI-powered scalping companion. From zero to trading in minutes.
Get trading in under 5 minutes with our streamlined setup process.
How to run the .exe file and initial system requirements.
MT5 credentials, parameters, and risk management settings.
EMA, HUNT, and AI Ensemble — choose your strategy.
Understanding the neural networks behind your trades.
Machine binding, validation, and account protection.
TP/SL settings, partial closes, trailing stops, and breakeven.
Your personal trading analyst powered by Llama 3.3 70B.
Common issues, error codes, and solutions.
From download to first trade in 4 simple steps
Double-click the MT5_LiqBot_AI_v10.exe file. No installation required — the bot runs as a portable application.
✅ Expected Result
The GUI window opens with dark theme, showing Parameters panel on the left, Live Chart in center, and AI Brain panel on the right.
📁 File Location
C:\Trading\Liqbot\MT5_LiqBot_AI_v10.exe
💾 System Requirements
Windows 10/11, 4GB RAM, MT5 installed
In the Parameters panel (top-left), fill in your MetaTrader 5 account details:
Your account number
Hidden for security
Broker server name
💡 Pro Tip: Use a Demo account first to familiarize yourself with the bot's behavior before trading real capital.
Scroll down to the License panel and enter your license key:
⚠️ Important: Each license is bound to ONE machine. Attempting to use the same key on another computer will result in a "BOUND_OTHER" error.
Click the START BOT button and watch the magic happen:
Launches the trading engine and connects to MT5
Halts all trading activities immediately
🎉 Congratulations!
Your bot is now running. Watch the AI Brain panel for real-time decisions, the Live Chart for market visualization, and Logs for activity tracking.
The Liqbot AI v10 interface is divided into 6 draggable, resizable panels. Each serves a specific function in your trading workflow.
Your command center. Configure MT5 credentials, trading parameters, risk settings, and select your trading mode.
Real-time price visualization with mode-specific overlays. EMA lines, liquidity zones, or AI ensemble indicators.
Timestamped activity log. Connection status, trades executed, errors, and system messages.
The neural network's thought process. See ensemble predictions, RL agent decisions, LSTM forecasts, and volatility regime.
History of completed trades with P&L, exit reasons, and performance metrics. Auto-exports to CSV.
Conversational AI powered by Llama 3.3 70B. Ask market questions, get TP/SL suggestions, or analyze your open positions.
All panels are draggable and resizable
Master every setting to optimize your trading performance.
| Setting | Description | Example | Impact |
|---|---|---|---|
| MT5 Login | Your MetaTrader 5 account number | 52558336 | Critical — Required to connect |
| MT5 Password | Your account password (hidden) | •••••••• | Critical — Authentication |
| MT5 Server | Broker server name exactly as in MT5 | ICMarketsEU-Demo | Critical — Connection endpoint |
| Instrument | Trading symbol (must match MT5 exactly) | US30, XAUUSD, BTCUSD | High — Determines market |
Minimum lot size per trade. Used as starting size with Auto MM enabled.
Maximum lot size cap. Safety limit to prevent overexposure regardless of performance.
Increment size for Auto MM. Higher values = more aggressive compounding.
When Auto MM is enabled, the bot automatically adjusts lot size based on performance:
Hard stop for the trading day. When realized + unrealized P&L reaches this negative threshold, all positions close and bot stops.
Take profit for the session. When reached, bot closes all positions and stops — locking in gains.
Seconds between trades
Concurrent positions per direction
Prevents simultaneous BUY+SELL
Dynamically calculates Take Profit and Stop Loss based on Average True Range (ATR) — adapts to current market volatility.
Fixed point values for TP and SL. Same distance for every trade regardless of market conditions.
💡 Why Auto Mode? Forex pairs might need 50-100 points while Indices need 1000+ points. ATR automatically scales to the instrument's natural volatility, preventing stops that are too tight (whipsawed) or too loose (excessive risk).
Three distinct strategies, each optimized for different market conditions and trading styles.
Classic Trend Following
Price > EMA5 > EMA13 > EMA21 + Momentum = BUY
Liquidity Hunter
Multi-Model Intelligence
| Feature | EMA | HUNT | AI Ensemble |
|---|---|---|---|
| Best Market Condition | Strong Trends | Ranging/Swings | Any (Adaptive) |
| Trade Frequency | Moderate | High (Setups) | Selective (Quality) |
| Learning Curve | ⭐ Easy | ⭐⭐ Medium | ⭐⭐⭐ Advanced |
| Win Rate Potential | 55-65% | 60-70% | 65-75% (Mature) |
| Requires Training | ❌ No | ❌ No | ✅ Yes (30+ samples) |
| Recommended For | Beginners, Trend traders | Scalpers, Range traders | AI enthusiasts, All traders |
Understanding the neural architecture behind your trades
Three machine learning models vote on every trade decision:
Proximal Policy Optimization agent learns from trade outcomes:
Long Short-Term Memory network analyzes price sequences:
Processes last 10 bars to predict next movement direction. Implemented in pure NumPy — no PyTorch/TensorFlow required.
Model-Agnostic Meta-Learning adapts quickly to new instruments:
Order Flow Analysis detects bid/ask imbalances:
First 30 samples — AI observes and builds initial dataset. Relies heavily on base EMA signal (50% weight).
30-60 samples — Ensemble begins voting but with reduced confidence. Balanced weight between AI and traditional signals.
60+ samples — Full AI autonomy. Ensemble drives 75% of decision weight. High confidence predictions with multi-timeframe consensus.
Each license is cryptographically bound to a single machine. This prevents unauthorized sharing and ensures secure access.
First checks if license was validated within last 12 hours
Compares current hardware fingerprint with registered machine
POST request to metaquantuniverse.com with encrypted payload
⚠️ Warning: If you see "BOUND_OTHER" error, your license is already active on another machine. Contact support for transfer requests.
Your personal trading analyst powered by Llama 3.3 70B — integrated directly into the trading interface.
Based on your current US30 BUY position (0.1 lots @ 50157.4) and market volatility:
Risk/Reward: 1:0.5 | Consider trailing stop at 50160.0
The AI automatically sees:
Default key included. For personal Groq accounts:
✅ Bot is running when START BOT is grayed out and STOP BOT is active
⚠️ Emergency: Closing the window stops the bot but keeps MT5 positions open
Stop → Start without closing window
Close → Reopen application
Delete cache files → Restart
This error means your license key is already registered to a different computer. Each license is bound to one machine's hardware fingerprint. To transfer to a new computer, contact MetaQuantUniverse support with your license key and new Machine ID.
The AI requires approximately 60 labeled samples to reach the Mature phase. In passive learning mode (observing without trading), this takes roughly 2-3 hours of market data. With active trading, it can mature faster (30-40 trades). You'll see the phase indicator in the AI Brain panel change from LEARNING → GROWING → MATURE.
Currently, one license supports one active instance. However, you can switch between instruments without restarting — just change the Instrument field and the AI will adapt using MAML (Meta-Learning). For true multi-chart trading, multiple licenses are required.
The bot has a 12-hour license cache, so brief disconnections won't stop trading. However, real-time data from MT5 is required for signals. If MT5 disconnects, the bot will pause and retry connection every 5 seconds. Open positions remain managed by MT5's server-side stops even if the bot is disconnected.
Check these common causes:
The AI automatically saves to two files in the bot's directory: liqbot_v10_ensemble.pkl (machine learning models) and liqbot_v10_rl_agent.pkl (reinforcement learning agent). Copy these files to backup your AI's "brain." To restore, simply place them back in the bot folder before launching.
Recommendations by experience level:
Average True Range — volatility indicator used for dynamic TP/SL calculation.
Mandatory waiting period between trades to prevent overtrading.
Multiple ML models voting together for more accurate predictions.
Having simultaneous BUY and SELL positions. "No Hedge" prevents this.
Strategy targeting stop-loss clusters (liquidity) before reversal.
Model-Agnostic Meta-Learning — quick adaptation to new instruments.
Proximal Policy Optimization — reinforcement learning algorithm.
Difference between expected and actual execution price.
Classification of current market volatility (LOW/MEDIUM/HIGH/EXTREME).