Clash Chronicles Every Clash Royale deck tells a story—Clash Chronicles brings that story to life. Users simply upload a screenshot of their Clash Royale deck or battle screen. Behind the scenes, the application runs a custom-built computer vision pipeline that automatically detects and identifies the eight cards present in the image with high accuracy. The image is first preprocessed using OpenCV and Pillow before being passed through a PyTorch + OpenCLIP embedding pipeline. Each detected card is converted into a high-dimensional feature vector and matched against a FAISS vector database containing over 180 Clash Royale cards, allowing the application to recognize the player's exact deck in milliseconds. Once the deck has been identified, the detected characters are sent to DeepSeek V4 Pro through the Fireworks AI API. A carefully engineered prompt, enriched with character personalities, lore, and gameplay context, enables the model to generate an original three-act cinematic narrative tailored specifically to that deck. Stories are streamed to the user in real time, creating an engaging reading experience as each act unfolds. The web interface is designed with a medieval-inspired glassmorphism aesthetic, complete with smooth animations, responsive layouts, and automatic pagination that presents the story act-by-act while the AI continues generating the narrative. Clash Chronicles combines computer vision, vector search, and large language models to transform a simple game screenshot into a personalized fantasy adventure, giving every Clash Royale player a unique piece of lore starring their favorite deck.
Category tags: