---
title: Build for Production-Grade AI Search
subtitle: Engineered for real-time retrieval with the speed, accuracy, and scale that modern AI demands.
pill: WHY QDRANT?
startFree:
  text: Start Free in Qdrant Cloud
  url: https://cloud.qdrant.io/signup
featureCards:
  - id: metadata-filters
    title: Expansive Metadata Filters
    description: "Store metadata in JSON and use advanced filters, such as <code>nested</code>, <code>text</code>, <code>geo</code>, <code>has_vector</code>, and more."
    image:
      src: /img/home/metadata-filters.png
      alt: Metadata filters code illustration
    link:
      text: Learn About Metadata Filters
      url: /documentation/manage-data/payload/
    size: large
  - id: hybrid-search
    title: Native Hybrid Search (Dense + Sparse)
    description: Blend keyword and vector search in one query – use dense or sparse vectors. Supports BM25, SPLADE++, and miniCOIL.
    image:
      src: /img/home/native-hybrid-search.png
      alt: Native hybrid search illustration
    link:
      text: Explore Hybrid Search
      url: /documentation/search/hybrid-queries/
    size: large
  - id: multivector
    title: Built-in Multivector
    description: Set new standards for relevance; make the retrieval layer more expressive, flexible, and multimodal with multiple vectors per object.
    image:
      src: /img/home/multivector.png
      alt: Multivector illustration
    link:
      text: See Documentation
      url: /documentation/tutorials-search-engineering/using-multivector-representations/
    size: small
  - id: one-stage-filtering
    title: Efficient, One-Stage Filtering
    description: Filters are applied during HNSW traversal — no pre- or post-filtering. High recall with low latency, even under complex conditions.
    image:
      src: /img/home/one-stage-filtering.png
      alt: One-stage filtering illustration
    link:
      text: See Documentation
      url: /articles/filterable-hnsw/
    size: small
  - id: reranking
    title: Full-Spectrum Reranking
    description: Infuse business logic with score boosting, achieve token-level precision with late interaction models (e.g. ColBERT), diversify results with Maximum Marginal Relevance (MMR)
    image:
      src: /img/home/reranking.png
      alt: Reranking illustration
    link:
      text: See Documentation
      url: /documentation/tutorials-basics/reranking-hybrid-search/
    size: small
sitemapExclude: true
---