Vector databases are storage systems for multi-dimensional vectors that offer searches based on similarity and contextual meaning of data, providing a better fit for unstructured datasets such as video, audio, or text. Unlike traditional databases, vector databases can effectively handle big data without bottlenecks or latency issues. Algorithms like Nearest Neighbor Search, Approximate Nearest Neighbor Search, FAISS, SPTAG, and HNSW are used to search vector databases working with various types of data. Studies on vector databases provide a long history of representing data digitally by vectors, from classical to state-of-the-art word representation language models, and address perspectives in the context of Large Language Models. Despite being superior for finding relevance, vector search still has limitations such as scalability issues and costs. To improve outcomes, approaches like filtering metadata, hybrid search, re-ranking, and VectorShift are used to enhance vector search capabilities.