Stylized image generation from text prompts is an emerging area at the intersection of art and artificial intelligence, enabling users to convert natural language descriptions into visually compelling, artistically enriched imagery. The proposed approach presents a modular framework that combines semantic understanding of text with image synthesis using Stable Diffusion XL (SDXL), followed by a flexible stylization phase based on neural style transfer techniques. Unlike end-to-end pipelines, our method decouples image generation from artistic styling, allowing rapid re-stylization without the need to regenerate base images. Using OpenCV-based filters, styles such as pencil sketches are applied efficiently while preserving edges, tones, and overall structural clarity. Experimental results show that the system can produce visually rich, semantically aligned, and stylistically consistent outputs across various historical and conceptual prompts. This modularity also improves adaptability for diverse use cases. The proposed approach has potential in applications such as educational tools, digital heritage preservation, virtual museum curation, and AI-assisted storytelling, where both interpretability and creativity are essential.