nav.groups.storage
Backend: Qdrant
Qdrant embedding store backend for approximate nearest-neighbour search.
Backend: Qdrant
Qdrant embedding store.
Feature qdrant. Provides an EmbeddingStore implementation backed by Qdrant's vector database with HNSW-based approximate nearest-neighbour search.
Configuration
[store.embedding.qdrant]
url = "http://localhost:6334"
collection = "behest_embeddings"
vector_size = 1536
When to use
- More than ~10k embedding vectors (memory backend is brute-force).
- Production RAG with low-latency search.
- Existing Qdrant infrastructure.
See also
- EmbeddingStore — the trait.
- RAG Context Adapter — the consumer.