zilliztech/knowhere

Vector search engine inside Milvus, integrating FAISS, HNSW, DiskANN.

What it solves

Knowhere is the internal core of the Milvus vector database, providing the fundamental low-level operations required for high-performance vector search and indexing.

How it works

Written in C++, Knowhere acts as a standalone library that can be integrated into larger systems like Milvus. It supports multiple hardware targets, including CPU and GPU (via cuVS), and provides Python bindings through a Python wheel for easier integration with Python-based workflows.

Who it’s for

It is primarily intended for developers building vector databases or those who need a high-performance C++ core for vector similarity search and indexing.

Highlights

  • Independent C++ project serving as the core for Milvus.
  • Supports both CPU and GPU acceleration.
  • Provides Python bindings via SWIG.
  • Compatible with various Linux distributions and MacOS.

Related

  • Project
    milvus-io/milvus

    Milvus is a high-performance, distributed vector database built for scaling AI applications that require efficient search and organization of massive amounts of unstructured data.

  • Project
    milvus-io/pymilvus

    A Python SDK for Milvus, providing a programmatic interface to manage and search high-dimensional vector embeddings in the Milvus vector database.

  • Project
    milvus-io/milvus-lite

    A lightweight, pure-Python local version of the Milvus vector database designed for development, testing, and small-scale AI applications.

  • Project
    milvus-io/milvus-sdk-java

    A Java client library for Milvus, enabling Java developers to integrate vector database capabilities into their AI applications.

  • Project
    ztxz16/fastllm

    A high-performance C++ LLM inference engine that enables hybrid deployment of dense and MoE models across GPU, CPU, NUMA, and disk to maximize hardware utilization.