Vald is a fully scalable distributed vector search engine that is cloud-native. It leverages the fastest ANN algorithm NGT for searching neighbors and features automatic vector indexing and index backup. The search engine supports horizontal scaling designed for searching billions of feature vector data points.

The easy-to-use, highly customizable product was developed in GO, gRPC, Docker, Nodejs, and supports distributed vector spaces, containerization, microservices, real-time monitoring, and data persistency.

Vald runs on Kubernetes and Helm package manager helps users deploy and configure Vald in different environments. The NGT component guarantees high performance of the vector search engine.

Project Background

  • Database: Vald Vector Search Engine
  • Author: Vald
  • Initial Release: 2021
  • Type: Vector Search Engine
  • License: Apache License 2.0
  • Contains: Vald Agent, Vald LB Gateway, Vald Discoverer Vald Index Manager
  • Language: Golang, Java, Nodejs and Python
  • GitHub: /vdaas/vald has 811 stars, 79 releases, 16 contributors 
  • Runs On: Kubernetes, Go, Helm, libhdf5
  • Twitter: /vdaas_vald/ 


  • Asynchronize Auto Indexing
  • Customizable Ingress/Egress Filtering
  • Distributed Indexing
  • Index replication
  • Auto Indexing Backup 
Scroll to Top