This project forecasts exchange-rate volatility in high-remittance economies by combining macroeconomic indicators with news-based text analysis. Using ML models (logistic, random forest, GBM) and text sentiment, we identify early FX stress signals and show macro factors dominate remittances.