Engine Overview

Updated at:
Copy as MD

The Lindorm vector engine stores, indexes, and retrieves vector data at scale. It supports multiple indexing algorithms, distance functions, and a variety of integrated data retrieval methods, including integrated full-text and vector retrieval—essential for Retrieval-Augmented Generation (RAG) systems to improve the accuracy of large models. Use cases include personalized recommendation, NLP, and intelligent Q&A.

Key features

  • Low cost and high performance

    Disk-based indexing and shared storage architecture enable cost-effective storage. Supports tens of billions of vectors per index with query latency in tens of milliseconds.

  • Ease of use

    Supports real-time data updates and access over protocols such as OpenSearch, SQL, and REST.

  • Multimodal capabilities

    Combines scalar, full-text, and vector retrieval for flexible multimodal queries and various data query capabilities.

  • One-stop solution

    Integrates with the built-in AI engine for embedding inference, providing foundational capabilities for large-model RAG systems.