Tools · SPBoucher
Menu

Qdrant

In-depth entry

by Qdrant Solutions GmbH · Database

Open-source vector search engine written in Rust with filterable HNSW indexes, sparse and multi-vector support, REST/gRPC APIs and a managed cloud

Apache-2.0 licenseOpen sourceFreemiumUsage-basedEnterpriseSelf-hostedCloudLocalSelf-hostable
Visit official site Compare

Link verified Nov 2, 2026

Overview

Qdrant is an open-source vector database and similarity search engine written in Rust by Qdrant Solutions GmbH and released under the Apache 2.0 license. It stores points consisting of one or more vectors plus a JSON payload, builds HNSW indexes over them, and answers nearest-neighbor queries that can be constrained by payload filters evaluated during the graph search rather than after it. Dense, sparse and multi-vector (late-interaction) representations can coexist in one collection, which supports hybrid lexical-plus-semantic retrieval and reranking pipelines.

The server exposes REST and gRPC APIs with official clients for Python, JavaScript/TypeScript, Rust, Go, Java and .NET, and it integrates with LangChain, LlamaIndex, Haystack and most embedding providers. Scalar, product and binary quantization reduce memory, and on-disk payload and vector storage allow collections larger than RAM. Distributed mode shards and replicates collections across nodes with Raft-based consensus; snapshots cover backup and migration. Qdrant Cloud offers managed clusters with a free tier, and a hybrid option runs the data plane in the customer's own infrastructure.

Qdrant is a common choice for retrieval-augmented generation, semantic search, recommendation and deduplication where metadata filtering matters. It does not compute embeddings itself (apart from the optional FastEmbed helper in client libraries), has no SQL interface or joins, and filter-heavy queries need payload indexes to stay fast. Operating a multi-node cluster requires attention to shard placement and memory, and the single-binary deployment on macOS or Linux is often sufficient for research-scale collections.

Capabilities

  • HNSW approximate nearest-neighbor search with in-graph payload filtering
  • Dense, sparse and multi-vector named vectors per point
  • Scalar, product and binary quantization
  • REST and gRPC APIs with clients in six languages
  • Distributed sharding and replication with Raft consensus
  • Snapshots and on-disk storage for collections larger than memory
  • Hybrid search and server-side fusion/reranking queries

Typical workflows

  1. Vector store for a RAG pipeline built with LangChain or LlamaIndex
  2. Hybrid keyword and semantic search over document collections
  3. Recommendation and deduplication with payload filters
  4. Single-binary local deployment on an Apple Silicon workstation

Strengths

  • Apache 2.0 license and self-hostable single binary
  • Filtering integrated with the vector index
  • Memory reduction via quantization
  • Managed cloud with free tier

Limitations

  • No embedding generation on the server
  • No SQL or relational joins
  • Payload filters need indexes to avoid slow scans
  • Cluster operations add complexity compared with pgvector inside Postgres

Prerequisites

  • Docker or a native binary, or a Qdrant Cloud account
  • An embedding model or API to produce vectors

Who it is for

  • Developers building RAG and semantic search
  • ML engineers running retrieval experiments

Alternatives & comparison

pgvector keeps vectors inside PostgreSQL with SQL and transactions but without a dedicated distributed index service. Weaviate bundles vectorizer modules and a GraphQL API; Milvus targets very large distributed deployments with many index types; Chroma emphasizes a minimal developer API for prototypes. Pinecone is proprietary and cloud-only. Qdrant's distinguishing traits are its Rust implementation, filterable HNSW and quantization options under a permissive license.

  • Chroma

    Open-source embedding database with a minimal Python and JavaScript API, usable in-process, as a server or through Chroma Cloud

    Open sourceOpen source
  • Elasticsearch

    Distributed search and analytics engine on Apache Lucene with full-text, aggregations, vector search and ES|QL; triple-licensed AGPL, SSPL and Elastic License

    MixedOpen source
  • Meilisearch

    Rust search engine for application search-as-you-type with typo tolerance, faceting and hybrid semantic search; MIT core plus BSL enterprise parts

    MixedOpen source
  • Milvus

    Open-source vector database from Zilliz with Lite, Standalone and Distributed deployments, many index types and GPU acceleration

    Open sourceOpen source
  • OpenSearch

    Apache-2.0 search and analytics suite forked from Elasticsearch and Kibana, governed by the OpenSearch Software Foundation

    Open sourceOpen source
  • pgvector

    PostgreSQL extension adding vector types, distance operators and HNSW/IVFFlat indexes for similarity search inside the database

    Open sourceOpen source
  • Pinecone

    Fully managed, proprietary vector database service with serverless indexes, namespaces, metadata filtering and integrated embedding and reranking

    ProprietaryFreemium
  • Typesense

    GPL-licensed in-memory search engine in C++ offering typo-tolerant instant search, faceting, vector and hybrid search, with Typesense Cloud hosting

    Open sourceOpen source
  • Weaviate

    Open-source vector database in Go with pluggable vectorizer modules, hybrid BM25 and vector search, multi-tenancy and a managed cloud

    MixedOpen source
Compare Qdrant with its alternatives →

Sources & verification

Primary sources consulted for this record. Claims are tied to the fields each source supports; the automated link checker re-verifies every URL on a schedule.

  • License
    Qdrant LICENSE (Apache 2.0)

    https://github.com/qdrant/qdrant/blob/master/LICENSE

    Supports: license_spdx, source_model

    checked Nov 1, 2026
  • Official documentation
    Qdrant Documentation

    https://qdrant.tech/documentation

    Supports: capabilities, interfaces, languages

    checked Nov 1, 2026
  • Official site
    Qdrant – Vector Database

    https://qdrant.tech/

    Supports: name, summary

    checked Nov 1, 2026
  • Pricing page
    Qdrant Cloud pricing

    https://qdrant.tech/pricing

    Supports: pricing_model

    checked Nov 1, 2026
  • Repository
    qdrant/qdrant

    https://github.com/qdrant/qdrant

    Supports: repo_url

    checked Nov 1, 2026
Confidence
High
Reviewed
Nov 1, 2026
Links
Verified · Nov 2, 2026

Spotted an error?

Broken link, outdated license, wrong category, or a tool that no longer exists? Tell us; reports are reviewed by the editor.

Meilisearch

Meili SAS · Vector databases

Rust search engine for application search-as-you-type with typo tolerance, faceting and hybrid semantic search; MIT core plus BSL enterprise parts

MIT AND BUSL-1.1Open sourceFreeSelf-hostedNov 2, 2026

OpenSearch

OpenSearch Software Foundation (Linux Foundation) · Vector databases

Apache-2.0 search and analytics suite forked from Elasticsearch and Kibana, governed by the OpenSearch Software Foundation

Apache-2.0Open sourceFreeSelf-hostedBroken

Typesense

Typesense, Inc. · Vector databases

GPL-licensed in-memory search engine in C++ offering typo-tolerant instant search, faceting, vector and hybrid search, with Typesense Cloud hosting

GPL-3.0Open sourceFreeSelf-hostedNov 2, 2026