Qdrant
App : A product you use directly
Open source vector database written in Rust: it stores your documents as vectors and retrieves the closest passages by meaning, the search building block of a RAG system.
The European choice for RAG or agent memory: editor based in Berlin, Apache 2.0 licence, deployable on your own infrastructure via Docker up to isolated environments, or as a managed cloud service with a small free cluster for testing. A minimum level of technical proficiency is required; for a few hundred documents, the vector extension of your PostgreSQL database may be enough.
Security reflects where data goes, requested permissions and incidents on record at the verification date. It is not a guarantee of zero risk: check the deployment mode and the connections you enable.
What it does, concretely
When an application needs to retrieve relevant passages from a large document collection, classic keyword search often misses the actual meaning of sentences. Qdrant solves this by storing each document as a numerical vector and finding the closest entries by meaning, not just by exact terms.
In practice, you launch Qdrant via Docker or connect to Qdrant Cloud, then send your vectors (produced by an external embedding model or via Qdrant Cloud Inference) through a REST or gRPC API. Qdrant indexes these vectors using the HNSW algorithm, applies filters on JSON metadata during index traversal, and returns the closest results in milliseconds. A built-in web UI lets you explore collections and test queries without writing code.
Qdrant is aimed at developers and technical teams building RAG (retrieval-augmented generation) systems, AI agents, semantic search engines, or recommendation systems. The tool offers official client libraries in Python, JavaScript/TypeScript, Go, Rust, Java, and .NET, as well as a lightweight version called Qdrant Edge for embedded devices or offline environments.
However, Qdrant is not suited for someone looking for a no-code or non-developer solution. For a few hundred documents, the vector extension of an existing PostgreSQL database may be sufficient and requires less configuration.
What it offers
- ·Dense, sparse, and multi-vector search: support for semantic embeddings, full-text search (BM25, SPLADE++), and late interaction models like ColBERT.
- ·Hybrid search: combining dense and sparse vectors in a single query, with result fusion via RRF or DBSF.
- ·JSON metadata filtering: advanced filters (text, geolocation, numeric ranges, nesting) applied during HNSW traversal without pre- or post-filtering.
- ·Vector quantization: memory usage reduction of up to 64x using asymmetric, scalar, and binary quantization.
- ·Built-in web UI: explore collections, test queries, and inspect results without writing code.
- ·Qdrant Edge: lightweight embedded version for edge devices, robots, or offline environments, with optional synchronization to a Qdrant server.
- ·Distributed deployment: horizontal scaling via sharding and replication with zero downtime.
- ·Qdrant Cloud Inference: generate text and image embeddings directly within Qdrant Cloud, no separate pipeline needed. (payant)
In practice
- Price
- Freemium
- Commercial use
- Commercial use: yes
- Available on
- cli, api, cloud
- Licence
- Apache 2.0
- FR interface
- No
Security and data
- Rating
- Security: one caveat, see the card
- Your data
- Local or cloud depending on use
- Maintenance
- Active
Links and repository checked on 29 October 2026 · added on 29 October 2026
How to install Qdrant
docker run -p 6333:6333 qdrant/qdrant