Weaviate, a vector database with ANN Index and CRUD support

The demand for vector similarity search solutions has increased due to the rising popularity of machine learning models. Approximate Nearest Neighbor (ANN) models like Annoy, faiss, and ScaNN have been developed to enable fast and large-scale vector searches by making a trade-off between accuracy and retrieval speed. However, many ANN libraries fall short in terms of real-time capabilities, mutability, persistence, consistency, resiliency, and scalability. Weaviate is an open-source, cloud-native, modular, real-time vector database that aims to combine the speed and large-scale capabilities of ANN models with all the functionality we enjoy about databases. It supports full CRUD capabilities, mutability of its indexes, and persistence of every single write. Weaviate is built around the idea of modularity and uses a custom HNSW implementation to overcome the limitations of existing HNSW libraries.