Epsilla is an open-source vector database that focuses on scalability, high performance, and cost-effectiveness in vector search. It serves as the Hippocampus for AI, bridging the gap between information retrieval and memory retention in Large Language Models. Noteworthy features include high performance and production-scale similarity search for embedding vectors, a database management system with familiar concepts, native Python support, and a REST API interface. Epsilla’s core is written in C++ and employs advanced academic parallel graph traversal techniques for vector indexing, resulting in 10 times faster vector search than HNSW while maintaining precision levels of over 99.9%. To get started, users can follow the provided steps to run the backend in Docker and interact with the Python client.
https://github.com/epsilla-cloud/vectordb