TECHNOLOGY PROFILE · TECH-0221
Qdrant
Vector database with payload filtering and retrieval APIs.
Reference link only — not reviewed this edition. This profile preserves the prepared catalog’s original descriptions. Advantages depend on the workload; comparisons are editorial, not benchmark results.
Where it has an edge
Combines vector search with metadata-aware filtering and service operations beyond an in-process ANN library.
Best-fit problems
Production semantic search with structured filters.
Tradeoffs & cautions
Adds database operations; compare filter behavior, recall and memory on real data.
What to test before choosing
Measure recall@k, filtered recall, indexing cost, p95 query latency, updates, backup and access control.
Runtime & hardware
Embedded, self-hosted or managed
Access & licensing
Open-source or commercial depending on edition
Execution
Local or hosted
Lifecycle
Not independently audited
Comparison basis
Editorial problem-fit judgement; no head-to-head benchmark performed
Connected concepts
Alternatives to investigate
pgvector · Milvus · Weaviate