Skip to main content

Qdrant Solutions GmbHVector Database

Qdrant

Vector database for dense, sparse and multivector search with JSON payload filters, hybrid search and built-in quantization. Self-hosted under the Apache License 2.0, or Qdrant Cloud on AWS, GCP and Azure with a free tier and usage-based Standard billing.

The 30-Second Take

Pick Qdrant if your retrieval needs its own store and leans on metadata filters.

Filtered search turns fast once every filtered field is indexed. Manfred Dreese reports P99s falling from hundreds of milliseconds to single digits.

Best fit
AI and platform teams whose retrieval needs a dedicated store with selective metadata filters, self-hosted or on Qdrant Cloud.
Main constraint
Memory and disk needs run high, and patch releases have regressed, so clusters need headroom and staged upgrades.
First move
Create payload indexes for every field you filter on before ingesting.

Why

Plan memory and disk headroom up front, sized on your own data. In Qdrant, deletes do not free disk until the optimizer runs.

Practitioner sentiment, teams and leaders−40 Negative

40 practitioners, 72 sources reviewed, Updated Nov 1, 2026How this is measured

Read the GILD Review
Save a product your organization uses.Keep up with Qdrant without adding it to your stack.Visit Website
Operating stack
AI Application Platform

Comparisons

How does Qdrant compare with the alternatives?

  1. Qdrant vs Pinecone

    • Vector Databases

    Qdrant also runs on your own servers and ARM machines, while Pinecone runs only as a managed service, in your cloud on Enterprise. Both start free, and Pinecone's Standard and Enterprise plans carry monthly minimums.

  2. Qdrant vs pgvector

    • Vector Databases

    pgvector keeps vectors in the Postgres you already run, while Qdrant runs as its own database, self-hosted or on Qdrant Cloud. Qdrant indexes filters into its search, and pgvector filters after the index scan, with iterative scans from 0.8.0 to fetch more rows.

Operator Insights

What are teams learning from building on Qdrant and running it in production?

See all 4 operator accounts

Buying and deployment

What a team building on it should verify before production traffic grows.

Sourced from Qdrant documentation

How you buy it

Free, then usage-based

Apache 2.0 when self-hosted. Cloud Standard is usage-based, and Premium is a committed spend.

Source 1, Qdrant, Pricing: PricingSource 2, Qdrant, README: READMESource 6, Qdrant, Premium tier: Premium tierSource 5, Qdrant, Pricing and payments: Pricing and paymentsSource 3, raw.githubusercontent.com, LICENSE: LICENSE

Details

Self-hosted Qdrant is licensed under the Apache License, Version 2.0, a no-charge, royalty-free license. Qdrant Cloud's Free Tier is free forever: one 0.5 vCPU, 1 GB RAM, 4 GB disk node, no credit card needed. Standard is usage-based, paid by credit card or an AWS, GCP or Azure Marketplace subscription. Premium is a committed-usage plan with a minimum spend agreed with Qdrant. Hybrid Cloud is part of the Enterprise plan, and Hybrid and Private Cloud start by contacting Qdrant.

Cost drivers

Billed hourly

Compute, memory, storage, backups and paid inference tokens after the monthly free allowance.

Source 1, Qdrant, Pricing: PricingSource 12, Qdrant, Cloud Inference: Cloud InferenceSource 5, Qdrant, Pricing and payments: Pricing and paymentsSource 2, Qdrant, README: README

Details

Qdrant Cloud Standard has usage-based pricing. Billing follows actual resource usage of compute (vCPU), memory (GB), cluster storage (GB), backup storage (GB) and used inference tokens of paid models, and usage is billed hourly. Cloud Inference is billed per token at a fixed, region-independent price that depends on the model, shown in the Cloud Console. On paid clusters, paid models run on metered tokens after a monthly free allowance of up to 5 million tokens per model.

Where it runs

Self-hosted or Cloud

Self-hosted needs block storage and does not work with NFS or S3. Cloud runs on AWS, GCP and Azure.

Source 20, Qdrant, Installation: InstallationSource 21, Qdrant, Managed Cloud: Managed CloudSource 4, Qdrant, Create a cluster: Create a clusterSource 5, Qdrant, Pricing and payments: Pricing and paymentsSource 22, Qdrant, Security: Security

Details

Qdrant runs self-hosted on 64-bit x86_64 or ARM64 systems and needs block-level storage with a POSIX-compatible file system. It does not work with NFS or S3. Qdrant Managed Cloud runs natively on AWS, GCP and Azure. On Qdrant Cloud, data in Qdrant clusters is stored only in the cluster's deployment region, and telemetry and logs are pushed to Qdrant's US management plane.

What to verify

SSO on Premium

Console sign-in only. Audit logs cover paid clusters from v1.17.0, and logs go to Qdrant's US plane.

Source 6, Qdrant, Premium tier: Premium tierSource 15, Qdrant, Cloud account setup: Cloud account setupSource 23, qdrant.tech, Configure a cluster: Configure a clusterSource 22, Qdrant, Security: Security

Details

Enterprise SSO needs the Premium tier and covers Cloud Console sign-in only. The security page calls SSO a Premium add-on, and the Premium docs list it as included. Audit logs are available for paid clusters from Qdrant v1.17.0. On Qdrant Cloud, telemetry and logs are pushed to Qdrant's US management plane. Unused free clusters are suspended after 1 week and deleted after 4 weeks of inactivity if not reactivated. Under section 10(1) of the Cloud Service Agreement, Qdrant may "use, copy, extract, modify, distribute, analyze, compile, and display" Aggregated Data, usage statistics that do not identify the customer, "for routine business purposes", including developing and improving Qdrant products and services, and may "share versions of" it.

Signals

What is the current evidence saying about Qdrant?

A current read of what teams report publicly about running Qdrant, self-hosted and on Qdrant Cloud, with each finding linked to its sources.

Whose sources count

Practitioner sentiment, teams and leaders

−40 Negative net

of 41 sources that take a side, each author counted once.

How this figure is made

Current sentiment

Positive, mixed and negative sources

The score counts them as 7 positive, 10 mixed and 23 negative votes, one per author.

Teams running Qdrant clusters report fast filtered retrieval once payload indexes are in place, and recurring work on memory sizing, upgrades and cluster recovery. Regressions in 1.16.1 and 1.18.3 needed a later release or a downgrade.

Run a pilot on your own documents and filters to measure recall against exact search, query latency at your filter selectivity, and memory per million vectors.

Evidence field. All 72 sources, one mark each.

Arrange by

Each mark is one source: positive, mixed, negative, no opinion stated, or documentation (not counted).

By sentiment, Teams and technical leaders. Every source GILD read, grouped as the field draws it.
GroupSourcesPositiveMixedNegativeNo opinion stated and not counted in the scoreDocumentation, checked and not counted
Positive880000
Mixed10010000
Negative23002300
Documentation24000024
Outside this group722300
Every source GILD read, by the quarter it was published, as the field's time axis counts it. Documentation is dated by the day it was checked.
QuarterSources
Q1 20255
Q2 20252
Q3 20257
Q4 20257
Q1 20266
Q2 20267
Q3 202611
Q4 20263
Documentation, checked24

By sentiment

24 documentation pages check facts and are not counted; 7 sit outside this group.

Hover a mark for a quick look, click it or press Enter to open its source.Tap a mark to open its source, then step to its neighbors with Previous and Next.

What practitioners praise

Each mark is a source on it; filled marks praise it.

  1. Low Operational Work6 of 7 sources on it praise it

Most common concerns

Top concern of 4. Each mark is a source on it; filled marks raise it.

  1. Cluster OperationsAll 9 sources on it raise it as a concern
All 72 sources behind these findings
  1. READMEQdrant (Official), Product documentation, checked Nov 1, 2026
  2. Embeddings: bring your own vectorsQdrant (Official), Product documentation, checked Nov 1, 2026
  3. InferenceQdrant (Official), Product documentation, checked Nov 1, 2026
  4. FastEmbedQdrant (Official), Product documentation, checked Nov 1, 2026
  5. Cloud account setupQdrant (Official), Product documentation, checked Nov 1, 2026
  6. LangChain integrationQdrant (Official), Product documentation, checked Nov 1, 2026
  7. Framework integrationsQdrant (Official), Product documentation, checked Nov 1, 2026
  8. Haystack integrationQdrant (Official), Product documentation, checked Nov 1, 2026
  9. InstallationQdrant (Official), Product documentation, checked Nov 1, 2026
  10. Managed CloudQdrant (Official), Product documentation, checked Nov 1, 2026
  11. SecurityQdrant (Official), Product documentation, checked Nov 1, 2026
  12. PricingQdrant (Official), Product documentation, checked Nov 1, 2026
  13. Cloud InferenceQdrant (Official), Product documentation, checked Nov 1, 2026
  14. Marketplace billingQdrant (Official), Product documentation, checked Nov 1, 2026
  15. Premium tierQdrant (Official), Product documentation, checked Nov 1, 2026
  16. Self-hosted audit loggingQdrant (Official), Product documentation, checked Nov 1, 2026
  17. GDPR data processing agreementQdrant (Official), Product documentation, checked Nov 1, 2026
  18. Cloud Service AgreementQdrant (Official), Product documentation, checked Nov 1, 2026
  19. Cloud backups: retentionQdrant (Official), Product documentation, checked Nov 1, 2026
  20. Create a clusterQdrant (Official), Product documentation, checked Nov 1, 2026
  21. VectorsQdrant (Official), Product documentation, checked Nov 1, 2026
  22. Cloud Inference docsQdrant (Official), Product documentation, checked Nov 1, 2026
  23. Qdrant Cloud SLAQdrant (Official), Product documentation, checked Nov 1, 2026
  24. Resource quotasQdrant (Official), Product documentation, checked Nov 1, 2026
  25. GitHub issue by dstocktondstockton (David Stockton), GitHub issue reporter (self-described: A 3-node cluster after raising a collection's replication factor), GitHub issue, published Oct 23, 2026
  26. GitHub issue by rsicartR.Sicart (rsicart), JobTeaser (GitHub profile) (self-described: A 3-node Qdrant Hybrid Cloud cluster on Kubernetes with Calico eBPF), GitHub issue, published Oct 20, 2026
  27. GitHub issue by jkbeJakob Edding (jkbe), bakdata (GitHub profile) (self-described: A memory-scarce Qdrant deployment with only the HNSW index in memory), GitHub issue, published Oct 13, 2026
  28. GitHub pull request by doogle-ossDebabrata Patnaik (doogle-oss), Application Development Manager, JP Morgan Chase Global Payments (GitHub profile) (self-described: Approving FastEmbed for use inside a company), GitHub pull request, published Sep 19, 2026
  29. Reddit comment by Hungry-Horror-7577Hungry-Horror-7577, Reddit commenter (self-described: A fully offline RAG index over technical books on one machine), Reddit comment, published Sep 15, 2026
  30. PeerSpot review by Pawel CisloPawel Cislo, PeerSpot reviewer (self-described: MLOps engineer at a tech services company with 501 to 1,000 employees, per PeerSpot's reviewer line), PeerSpot review, published Aug 24, 2026
  31. GitHub issue by bvolkmerBenedikt Volkmer (bvolkmer), Machine Learning Researcher and Developer (GitHub profile) (self-described: Downloading snapshots through the Python client), GitHub issue, published Aug 19, 2026
  32. GitHub issue by scheung-lbgSebastian Cheung (scheung-lbg), LBG (GitHub profile) (self-described: Installing FastEmbed through a client's package firewall), GitHub issue, published Aug 18, 2026
  33. PeerSpot review by Uzair FaruqiUzair Faruqi, PeerSpot reviewer (self-described: Head of data science at Mjunction Services, per PeerSpot's reviewer line), PeerSpot review, published Aug 17, 2026
  34. GitHub issue by ronenperezronenperez, GitHub issue reporter (self-described: A production Qdrant cluster with on-disk vectors), GitHub issue, published Jul 27, 2026
  35. GitHub issue by cainzhongCain (cainzhong), Micro Focus (GitHub profile) (self-described: Scanning the latest qdrant/qdrant Docker image for vulnerabilities), GitHub issue, published Jul 17, 2026
  36. Hacker News comment by cbcoutinhoChris Coutinho (cbcoutinho), Data Engineer, Astrolabe Cloud (GitHub profile) (self-described: Building a self-hosted Nextcloud knowledge-base and agent-memory app), Hacker News comment, published Jul 16, 2026
  37. Reddit comment by tenequmtenequm, Reddit commenter (self-described: An app that needed a database with no running server), Reddit comment, published Jul 15, 2026
  38. Reddit comment by sinevilsonsinevilson, Reddit commenter (self-described: Running Qdrant in production with many collections), Reddit comment, published Jul 15, 2026
  39. Hacker News comment by vunderbaShaun Pedicini (vunderba), Hacker News commenter (HN profile) (self-described: Local multimodal image search on their own hardware), Hacker News comment, published Jun 24, 2026
  40. Reddit comment by Master_Page_116Master_Page_116, Reddit commenter (self-described: A RAG pipeline built on the Mastra agent framework), Reddit comment, published Jun 17, 2026
  41. Reddit post by SilverConsistent9222SilverConsistent9222, Reddit poster (self-described: An internal knowledge base tool built at work), Reddit post, published Jun 7, 2026
  42. PeerSpot review by Amrit DashAmrit Dash, PeerSpot reviewer (self-described: Automation engineer at an educational organization with 11 to 50 employees, per PeerSpot's reviewer line), PeerSpot review, published May 26, 2026
  43. Reddit comment by Zealousideal_Wolf624Zealousideal_Wolf624, Reddit commenter (self-described: Running Qdrant at scale), Reddit comment, published May 21, 2026
  44. PeerSpot review by Vedant WaniVedant Wani, PeerSpot reviewer (self-described: Analyst at Synergy Connect, per PeerSpot's reviewer line), PeerSpot review, published May 18, 2026
  45. Practitioner blog post by Manfred DreeseManfred Dreese, tech lead and consultant (his own blog) (self-described: Three years running Qdrant for a large-scale recommendation product on Kubernetes), Practitioner blog post, published Apr 23, 2026
  46. GitHub issue by kaancfidanKaan C. Fidan (kaancfidan), Vispera (GitHub profile) (self-described: A collection of about 30M vectors with nested array payloads), GitHub issue, published Mar 16, 2026
  47. Hacker News comment by arlogilbertArlo Gilbert (arlogilbert), serial founder (HN profile), Osano (GitHub) (self-described: Comparing Zvec, LanceDB and Qdrant on a small text dataset), Hacker News comment, published Feb 23, 2026
  48. GitHub issue by bytestreamKieran (bytestream), SupportPal (GitHub profile) (self-described: A staged upgrade from Qdrant 1.10.1 to 1.16.3), GitHub issue, published Feb 19, 2026
  49. Hacker News comment by sa-codeSamay Kapadia (sa-code), Principal ML Engineer, Zalando (GitHub profile linked from HN) (self-described: Setting a default vector store for a team), Hacker News comment, published Jan 25, 2026
  50. GitHub issue by 02denoDeniz Özdemirli (02deno), Computer Engineer at Nokia (GitHub profile) (self-described: Hybrid RRF queries with min_should filters on Qdrant 1.16.3), GitHub issue, published Jan 23, 2026
  51. Hacker News comment by p410n3p410n3, Hacker News commenter (self-described: Semantic product search for an e-commerce catalog), Hacker News comment, published Jan 9, 2026
  52. GitHub issue by ffletcherrffletcherr (Homayoun), GitHub issue reporter (self-described: A Kubernetes deployment with tens of millions of points), GitHub issue, published Dec 21, 2025
  53. GitHub issue by boris-petrovBoris Petrov (boris-petrov), profuz (GitHub profile) (self-described: Upgrading self-hosted Qdrant with an automated test suite), GitHub issue, published Nov 27, 2025
  54. GitHub issue by DragosGhineaDragos Dumitru Ghinea (DragosGhinea), AI Developer at It Just Works Inc. (GitHub profile) (self-described: Filtered search over a 5,657,847-point collection after upgrading to 1.16), GitHub issue, published Nov 25, 2025
  55. GitHub issue by avvertixAlessio (avvertix), Full-Stack Engineer and AI Engineer (GitHub profile) (self-described: Qdrant in Docker with data on a Windows-mounted volume), GitHub issue, published Nov 25, 2025
  56. Hacker News comment by couAUIAcouAUIA, Hacker News commenter (self-described: An engineering company building internal document search), Hacker News comment, published Nov 5, 2025
  57. GitHub issue by fraenkiFrank Wall (fraenki), markt.de (GitHub profile) (self-described: Operating a self-hosted Qdrant cluster), GitHub issue, published Oct 27, 2025
  58. GitHub issue by ApidcloudLuís Fernandes (Apidcloud), Artivive (GitHub profile) (self-described: Returning large payloads from Qdrant through the Node client), GitHub issue, published Oct 8, 2025
  59. Reddit comment by eujzmceujzmc, Reddit commenter (self-described: A SaaS team with millions of vectors choosing between Milvus and Qdrant), Reddit comment, published Sep 24, 2025
  60. Hacker News comment by k9294: Comparing Qdrant and Zilliz Cloud against turbopuffer on operational simplicityk9294, Hacker News commenter (self-described: Comparing Qdrant and Zilliz Cloud against turbopuffer on operational simplicity), Hacker News comment, published Sep 9, 2025
  61. GitHub issue by ml-abdulaml-abdula (Abdullah), GitHub issue reporter (self-described: A production deployment of about 4.6M points with vectors and HNSW on disk), GitHub issue, published Sep 3, 2025
  62. Reddit comment by junegirl_198junegirl_198, Reddit commenter (self-described: A production deployment with lots of ingests and lower query volume), Reddit comment, published Sep 3, 2025
  63. GitHub issue by eekRadu-Sebastian Amarie (eek), AI startup founder (GitHub profile) (self-described: Per-tenant semantic search over one large collection), GitHub issue, published Aug 26, 2025
  64. GitHub issue by orhanraufRauf Akdemir (orhanrauf), co-founder and CTO, Airweave (GitHub profile) (self-described: Self-hosting Qdrant v1.15.3 on Kubernetes (Azure AKS) with a 50,000-document collection), GitHub issue, published Aug 23, 2025
  65. GitHub issue by morsikKonrad Mosoń (morsik), DevOps, openhades (GitHub profile) (self-described: A 6-node Qdrant cluster), GitHub issue, published Aug 18, 2025
  66. GitHub issue by aalekhpatel07Aalekh Patel (aalekhpatel07), Canadian Centre for Child Protection (GitHub profile) (self-described: Restoring collections with aliases from snapshots), GitHub issue, published Jun 24, 2025
  67. GitHub issue by amircodotaamircodota, GitHub issue reporter (self-described: A large production collection on Qdrant 1.8.4), GitHub issue, published Jun 3, 2025
  68. GitHub issue by rajanpatilRajan Patil (rajanpatil), Elsevier (GitHub profile) (self-described: A single Qdrant node on Azure Kubernetes Service), GitHub issue, published Mar 28, 2025
  69. GitHub issue by bdeckelRobert Deckel (bdeckel), PTC (GitHub profile) (self-described: A single Qdrant pod on Kubernetes in Microsoft Azure with the default Helm values), GitHub issue, published Mar 21, 2025
  70. Hacker News comment by boboshabobosha, 3x entrepreneur and AI researcher (HN profile) (self-described: A team that ran Qdrant in production beside Postgres), Hacker News comment, published Mar 18, 2025
  71. Hacker News comment by k9294: A relatively small startup running managed Qdrantk9294, Hacker News commenter (self-described: A relatively small startup running managed Qdrant), Hacker News comment, published Mar 17, 2025
  72. GitHub issue by MohammadsakniniMohammad Saknini (Mohammadsaknini), Data & AI Engineer at Inform GmbH (GitHub profile) (self-described: Bulk ingestion of 89 million 768-dimension points in Docker on WSL, for research), GitHub issue, published Jan 8, 2025

Jan 2025 to Oct 2026.

48 sources from practitioners, by month and sentiment.

The full readExplore all Qdrant Signals6 praise themes, 5 concerns, 0 points of agreement and 1 disagreement across all three groups, with all 72 sources open to inspect.

Where It Fits

Where does Qdrant belong in the stack?

Open Where It Fits: the Stack Field and the full map

Qdrant is Core across Data, supports Operate, and touches Model, Build, and Evaluate.

Business area
Product & Engineering
Function
Software Development
Operating stack
AI Application Platform
Primary role
Vector Database
Qdrant beside the rest of its stackThe Stack Field, every product on every layer
FootprintQdrant across the AI Application Platform stack. Bar height is its role on each layer.

  1. Model

    1. Model APIs: Touches

      Model APIs, touches. Qdrant Cloud Inference hosts a range of embedding models, some for free, and can call embedding models from OpenAI, Cohere, Jina AI and OpenRouter with your own API key. Generation models stay with a model API.

    2. Model gateway: Touches

      Model gateway, touches. OpenRouter is one of the external embedding providers Cloud Inference can call with your own key. Routing generation traffic stays with a gateway.

  2. Build

    1. Orchestration: Touches

      Orchestration, touches. LangChain ships the langchain-qdrant partner package with dense, sparse and hybrid retrieval, the Qdrant team maintains the Haystack document store, and LlamaIndex is on Qdrant's framework list. The agent logic stays in the framework.

    2. Tool use: Touches

      Tool use, touches. Qdrant exposes a REST API with an OpenAPI 3.0 specification and a gRPC interface, with clients for Python, TypeScript, Rust, Go, Java and .NET, which an agent's retrieval tool calls.

  3. Data

    1. Retrieval: Core

      Retrieval, core. This is Qdrant's primary role: it stores dense, sparse and multivector embeddings, filters on JSON payloads, merges hybrid results with Reciprocal Rank Fusion or Distribution-Based Score Fusion, and cuts RAM with built-in quantization.

  4. Evaluate

    1. Evaluation: Touches

      Evaluation, touches. Qdrant serves the retrieval results an evaluation measures. Scoring answers stays with an evaluation product.

  5. Operate

    1. Observability: Touches

      Observability, touches. From Qdrant v1.17.0, audit logs record API operations on paid Cloud clusters and in self-hosted Qdrant. Tracing an application's requests stays with an observability tool.

    2. Cloud deployment: Supports

      Cloud deployment, supports. Qdrant Managed Cloud runs on AWS, GCP and Azure with data stored in the cluster's region, and Hybrid Cloud runs in your own infrastructure. The cloud account and the rest of the application stay with your cloud provider.

    3. Guardrails: Touches

      Guardrails, touches. Qdrant holds SOC 2 Type 2 and HIPAA certifications and offers single sign-on on Premium, which secure the data store. Prompt and output guardrails stay with a security product.

Market mapThe layers Qdrant covers, and who else owns each one.

  1. Retrieval & grounding

    CoreData

  2. Cloud deployment & capacity

    SupportsOperate

  3. Model APIs

    TouchesModel

  4. Model gateways & routing

    TouchesModel

  5. Orchestration & agent frameworks

    TouchesBuild

4 more layers on the full map. Core, Supports and Touches describe coverage. They say nothing about quality.

More on Qdrant

  1. GILD ReviewGILD's evaluation, with the buying facts
  2. Research1 piece
  3. Guides1 guide
  4. Operator Insights4 operator accounts
  5. Where It FitsOn the AI Application Platform stack
  6. Related Products4 products