kindofzero/forecasting-plus-mt5

Forecasting Plus MT5 Dashboard is a Streamlit based application that connects to MetaTrader 5, a trading platform, and adds AI powered price forecasting on top of it. It is aimed at people who trade on MT5 and want a visual dashboard that shows predicted price movement alongside live charts. The dashboard pulls in data for all the symbols available in a trader's MT5 account, sorted by category, and displays four timeframes at once: five minute, fifteen minute, one hour, and daily. Each chart shows a candlestick view combined with a forecast line and a confidence band, so a user can compare what actually happened against what the model expected. Under the hood, the project offers eight different deep learning forecasting models, including Kronos, ChronosX, Sundial, TiRex, MOIRAI-MoE, TimesFM, Time-MoE, and FinCast. These come from research groups such as Amazon, Google DeepMind, Salesforce, and IBM, and are pulled from HuggingFace Hub. Each model runs in its own isolated Python environment so that conflicting dependencies between models do not break the whole system. A setup script automatically builds these separate environments. If a model cannot load for any reason, the app falls back to a synthetic forecast, and that fallback state is clearly labeled in the interface so a user is not misled into thinking it is a real prediction. Beyond forecasting, the dashboard supports manual trading actions: setting lot size, setting take profit and stop loss in points, and sending buy or sell orders, plus closing existing positions for a chosen symbol. There is also an automated trading bot component that can act on signals without manual input. The README is explicit that buy, sell, and close order actions are real trades sent to the connected account, not simulated ones, so anyone using this tool is working with genuine money and market risk. Setup requires a Windows environment with MetaTrader 5 installed, a .env file with login credentials, a Python virtual environment, and the project's listed dependencies. This is not a beginner friendly plug and play tool: it assumes familiarity with MT5, Python environments, and the real risks of automated trading.