Qdrant
Rust vector search engine built for filtering-heavy AI workloads
Overview
Qdrant is an open-source (Apache-2.0) vector database and search engine written in Rust, with filtering during search, hybrid dense and sparse search, multivector and late interaction, and quantisation, available self-hosted or as Qdrant Cloud on AWS, GCP and Azure, plus hybrid and private cloud. Free forever cluster; Standard usage-based.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 1 Oct 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
- Single node: 0.5 vCPU, 1GB RAM, 4GB disk
- Free cloud inference with selected models
- Community support
- Suspended after 1 week unused
- Dedicated resources and scaling
- High availability, backup and DR
- 99.5% uptime SLA
- Support 10x5
- SSO
- Private VPC links
- 99.9-99.95% SLA
- 24x7 support
- Managed clusters on your infrastructure
- Air-gapped option
- Custom SLAs
Use cases
What you can produce with Qdrant
- Open-source vector database in Rust (Apache-2.0)
- Filtering during HNSW search
- Hybrid dense and sparse search (BM25, SPLADE, miniCOIL)
- Multivector, ColBERT and reranking
- Quantisation incl. 4-bit TurboQuant
- Qdrant Cloud, Hybrid and Private Cloud
Qdrant is one of the strongest open-source vector databases, especially where search has to be combined with complex filters, and its free cloud cluster and permissive licence make it easy to start. Its performance claims rest on its own benchmarks, and Cloud pricing is usage-based rather than a flat fee, so model costs before scaling. An excellent default for RAG and semantic search; compare with Milvus for very large deployments.
What we found
Qdrant is an open-source vector database and search engine, written in Rust, for retrieval-augmented generation, semantic search, recommendations and AI agents. It is known for filtering, which it applies during the search itself rather than afterwards, and supports native hybrid search combining dense vectors with BM25, SPLADE and its own miniCOIL sparse models, multivector and ColBERT-style late interaction, reranking, and quantisation to cut memory use — version 1.19, released in August 2026, added a 4-bit TurboQuant option. Qdrant publishes benchmarks showing it fast and memory-efficient; they are its own. It runs as a self-hosted engine under the Apache 2.0 licence or as Qdrant Cloud on AWS, Google Cloud and Azure, with hybrid and private-cloud options, a Qdrant Edge version for devices in beta and a serverless offering announced as coming soon.
Qdrant Solutions GmbH was founded in Berlin in 2021 by André Zayarni, its chief executive, and Andrey Vasnetsov, its technical chief. It raised a $7.5 million seed in 2023, a $28 million Series A led by Spark Capital in 2024 and a $50 million Series B led by AVP in March 2026, about $85.5 million in total. It reports more than 250 million downloads and over 100 staff in more than 20 countries, and names Canva, HubSpot, Roche, Bosch, Tripadvisor, Deutsche Telekom, Bazaarvoice and Flipkart among users.
The engine is free to self-host. Qdrant Cloud's free cluster, with no card required, is a single node with half a vCPU, 1GB of RAM and 4GB of disk, enough for about a million 768-dimension vectors according to its docs, and is suspended after a week unused. Standard is usage-based and billed hourly on CPU, memory, disk and backups, with high availability, scaling, backups, a 99.5 per cent SLA and weekday support. Premium, with a minimum spend, adds SSO, private links, up to a 99.95 per cent SLA and round-the-clock support, and Hybrid and Private Cloud deployments are quoted.
We could not read independent review sites. Qdrant suits developers who need fast, filter-heavy vector search, self-hosted or managed.
Compiled by RECATOOLS from public sources: vendor documentation and pricing, code repositories and published reporting.
About this listing
This entry was compiled from publicly available data including Qdrant's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with Qdrant unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to Qdrant directly →
Spotted something out of date? Suggest an update →
Related reading
Alternatives to Qdrant
More in Code & Dev Tools
