Endee.io Open-Sources High-Performance Vector Database
The company Endee.io has open-sourced its high-performance vector database. The move adds another competitor to the expanding landscape of vector databases for AI applications. This trend toward open-source and cloud-native entrants is increasing commoditization and putting pressure on established providers to deliver differentiated features.
- The open-source version of Endee is licensed under Apache 2.0, while a commercial enterprise version is offered by the parent company, Endee Labs Pvt Ltd. This enterprise edition includes features like hybrid search (combining vector and keyword search), advanced metadata filtering, and multi-tenancy. - Endee.io claims its database can achieve over 10,000 queries per second (QPS) with sub-5ms latency and over 99% recall, attributing this performance to its Vector Graph Engine (VGE). The VGE is described as an execution framework that utilizes layered memory, processor acceleration, adaptive caching, and intelligent quantization. - A key differentiator in the enterprise version is "Queryable Encryption," a feature that allows similarity searches to be performed directly on encrypted vectors. According to Endee.io, data is encrypted and decrypted on the client-side, meaning the server never has access to unencrypted data or queries, aiming for a zero-knowledge security model. - The company was founded by Vineet Dwivedi, a tech entrepreneur with a background in distributed systems and AI platforms. Endee was developed to address common issues with existing vector databases, such as high memory consumption and rising infrastructure costs. - Endee is positioned as a lightweight alternative to established vector databases like Pinecone, Qdrant, Milvus, and Weaviate, with a focus on reducing the hardware footprint required for high-performance AI applications. - The database is designed for a variety of AI workloads, including retrieval-augmented generation (RAG), semantic search, recommendation engines, and AI agents. It offers SDKs for Python, TypeScript, Go, and Java to facilitate integration.
