Combining computer vision with knowledge graphs, a recent Memgraph Community Call demonstrated how technologies like GraphRAG can transform images into queryable knowledge using a blend of computer vision, vector embeddings, and graph traversal. Led by Dino Duranovic and Ante Javor, the session highlighted the process of breaking down images into structured information and querying them semantically, inspired by human cognition. The project involves using neural networks to interpret images, CLIP for embedding text and images, and knowledge graphs for structuring visual data, enabling advanced search capabilities beyond traditional tags or metadata. The live demo featured a pipeline where images are transformed into a knowledge graph, facilitating both image and text queries, and emphasized the importance of graph databases over vector databases for contextual and expandable search results. The knowledge graph approach allows for exploring complex relationships and generating context-aware answers using large language models, showcasing the power of integrating multiple AI technologies for enhanced data interaction and retrieval.