Qdrant

license

Apache-2.0(OSI, engine core)

source

public(Rust)

managed-tier

Qdrant Cloud (hosted, paid + free tier)

core features withheld

no — one LICENSE file and no ee/, enterprise/ or commercial/ path at the repository root or among the lib/ crates. The pricing table marks Shard Splitting absent from open source and the documentation calls resharding Cloud-only, but the code is in lib/collection and turns on with the cluster.resharding_enabled setting, off by default, with no license check. GPU Indexing is marked as included in open source, so what is sold is operations around the engine

Qdrant's engine ships under Apache-2.0 from a repository with one license file and no enterprise directory. Resharding, which the pricing page and documentation reserve for Qdrant Cloud, is in the open source behind a configuration switch that is off by default, and what the paid tier otherwise adds is operations such as SLAs, managed backups and SSO.

• https://api.github.com/repos/qdrant/qdrant/contents recorded 2026-08-12

  Root tree: one LICENSE beside Cargo.toml, config, docs, lib, openapi, pkg, src, tests and tools. No ee/, enterprise/, commercial/ or proprietary/ path and no second license file, so no part of the engine ships under different terms.

• https://api.github.com/repos/qdrant/qdrant/contents/lib recorded 2026-08-12

  The engine's crates: api, blobstore, bm25, collection, common, edge, gpu, macros, posting_list, quantization, segment, shard, sparse, storage, trififo, uio-grpc-client, wal. GPU indexing is the gpu crate, published here, which is the first of the two features the pricing table marks as absent from OSS.

• https://github.com/qdrant/qdrant recorded 2026-06-04

  v1.18.2 (4 Jun 2026); Qdrant Cloud managed tier with free option

• https://github.com/qdrant/qdrant/blob/master/LICENSE recorded 2026-06-04

  Apache-2.0 license text

• https://qdrant.tech/documentation/scaling/distributed_deployment/ recorded 2026-09-28

  The distributed-deployment guide says Qdrant Cloud "offers the ability to automatically rebalance and to reshard your collections, which is not available in self-hosted Qdrant".

• https://qdrant.tech/pricing/ recorded 2026-09-28

  The pricing comparison marks "Uptime SLA", "Topology Aware Multi AZ Setup" and "Enterprise SSO Authentication" absent from open source and lists "Custom automation" for its backups. Of the engine rows, "GPU Indexing" is checked for open source and "Shard Splitting" is not; "Cloud Inference" is a hosted embedding service.

• https://raw.githubusercontent.com/qdrant/qdrant/master/lib/collection/src/collection/resharding.rs recorded 2026-09-28

  The collection-level resharding implementation, published in the open repository: it defines "pub struct AbortReshardingScope" and documents the consensus operations that "might trigger resharding abort". This is the shard-splitting mechanism the pricing table lists as absent from open source.

• https://raw.githubusercontent.com/qdrant/qdrant/master/lib/gpu/Cargo.toml recorded 2026-09-28

  The manifest of the lib/gpu crate, the engine's GPU indexing code, declares the crate Apache-2.0 with authors "Qdrant Team <info@qdrant.tech>", and its gpu feature pulls in the Vulkan bindings "dep:ash".

• https://raw.githubusercontent.com/qdrant/qdrant/master/LICENSE recorded 2026-09-28

  The master-branch LICENSE file is the full Apache License text, headed "Apache License" and "Version 2.0, January 2004".

• https://raw.githubusercontent.com/qdrant/qdrant/master/README.md recorded 2026-09-28

  The README states "Qdrant is licensed under the Apache License, Version 2.0", gives a "docker run -p 6333:6333 qdrant/qdrant" self-hosted start, and names the hosted service as an alternative: "Qdrant is also available as a fully managed" Qdrant Cloud with a free tier. Searched the body for enterprise, license key and commercial license: none.

• https://raw.githubusercontent.com/qdrant/qdrant/master/src/settings.rs recorded 2026-09-28

  The cluster settings carry resharding as a plain configuration flag, "pub resharding_enabled: bool, // disabled by default". Searched the file for license: none.

Qdrant is a server. Its Python client, qdrant-client, is left out because client installs can vary independently of server deployments. The score comes from Docker Hub pulls of the server image averaged over the image's lifetime, so for a product still growing it is a floor rather than a ceiling.

• https://hub.docker.com/v2/repositories/qdrant/qdrant/ recorded 2026-09-24

  "pull_count": 47050807 cumulative pulls of qdrant/qdrant, "date_registered": "2022-02-03", about 845,000 a month over the image's lifetime

• https://pypistats.org/api/packages/qdrant-client/recent recorded 2026-09-24

  "last_month":11277295 downloads of qdrant-client, the client package left out of the level

• https://ungh.cc/repos/qdrant/qdrant recorded 2026-09-24

  "stars":34774 for qdrant/qdrant. Corroboration only; a star count is not a use.

Qdrant is a complete, distributed vector engine comparable to Milvus. It fuses scores over vectors the caller supplies rather than hosting and evaluating models inside the serving path, as Vespa does.

• https://ann-benchmarks.com/ recorded 2026-09-30

  ANN-Benchmarks page headed "Benchmarks for Single Queries" links "Milvus(Knowhere)" and a "qdrant" algorithm entry among its results, with plots per dataset and no ranking table.

• https://qdrant.tech/benchmarks/ recorded 2026-09-30

  Vendor benchmark says Qdrant "achieves highest RPS and lowest latencies in almost all the scenarios" and its meta keywords include "Qdrant vs Milvus".

• https://qdrant.tech/documentation/concepts/hybrid-queries/ recorded 2026-09-30

  Hybrid queries page documents "Reciprocal Rank Fusion (RRF)" and "Distribution-Based Score Fusion (DBSF)" fusion plus a "Multivectors and Late Interaction" section.

• https://qdrant.tech/documentation/guides/quantization/ recorded 2026-09-30

  Quantization guide says users can "compress high-dimensional vectors, reduce memory footprint, and accelerate similarity search while preserving recall quality."

• https://raw.githubusercontent.com/qdrant/qdrant/master/README.md recorded 2026-09-30

  README says "Qdrant supports dense vectors for semantic similarity, sparse vectors for full-text search, and multivector search for objects with multiple embeddings or late interaction models like ColBERT." and "Scale horizontally with sharding and replication, and update or resize collections with zero downtime."