epsilla-cloud/vectordb

A high performance Vector Database leveraging parallel graph computing

What it solves

Epsilla is an open-source vector database designed to provide high-performance, scalable, and cost-effective similarity search for embedding vectors, bridging the gap between information retrieval and memory retention for Large Language Models.

How it works

Written in C++, the system uses advanced parallel graph traversal techniques for vector indexing. This allows it to achieve search speeds up to 10 times faster than HNSW while maintaining precision levels above 99.9%. It treats vectors as a standard field type within a full database management system, supporting compute-storage separation and a cloud-native, serverless architecture.

Who it’s for

Developers building AI applications that require production-scale vector search, hybrid search (dense and sparse vectors), and seamless integration with frameworks like LangChain and LlamaIndex.

Highlights

  • High Performance: 10x faster vector search than HNSW with over 99.9% precision.
  • Hybrid Search: Combines dense and sparse vectors for improved retrieval.
  • Built-in Embedding Support: Enables a "natural language in, natural language out" search experience.
  • Database Management: Familiar table and field concepts with metadata filtering.
  • Cloud Native: Supports multi-tenancy, serverless deployment, and compute-storage separation.
  • Broad Integration: Provides Python, JavaScript, and Ruby clients, and a REST API.

Related

  • Project
    gcpaas/DataRoom

    A low-code data visualization and dashboard design platform that supports AI-assisted page generation via the MCP protocol and integrates with over 20 different data sources.

  • Project
    annotorious/annotorious

    Annotorious is a lightweight JavaScript library for adding image annotation features to any website — install it, wire it to an image, and users can create, load, and track annotations with just a few lines of code. Open source (BSD 3-Clause) with free community support and paid consulting available.

  • Project
    LeslieLeung/glean

    A self-hosted RSS reader and personal knowledge management tool that uses intelligent aggregation and vector databases to help users manage high-volume reading.

  • Project
    bacalhau-project/bacalhau

    Bacalhau is a distributed compute orchestration framework that enables "Compute Over Data," allowing users to run jobs directly where large datasets are stored to reduce latency and costs.