The financial markets are extremely dynamic and non-linear, predicting stock values is a challenging undertaking. Conventional statistical techniques, such as linear regression and ARIMA, frequently fall short of capturing the intricate temporal dependencies and obscure connections that control market conditions. To overcome these constraints, this study suggests a dashboard framework driven by AI that smoothly combines sophisticated deep learning models with a system for user-centered visualization and interaction. The framework makes it possible to collect, preprocess, and normalize data from live sources like Yahoo Finance. This research prepares a comprehensive framework, paired with an API readily exposed to a graphical user-interface aimed at creating, training and comparing multiple neural network architectures over configurable parameter spaces, and a visual comparison to discern peculiarities and assess model performances over real data sourced live from the internet. In addition to providing opportunities for future developments in risk modeling and multimodal forecasting, the implementation shows great promise for real-world deployment in financial technology ecosystems.