TECHNOLOGY PROFILE · TECH-0072
FAISS
Library for dense-vector similarity search and clustering.
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
In-process CPU/GPU indexes avoid database-service overhead when only vector search is needed.
Best-fit problems
Large local embedding indexes and retrieval research.
Tradeoffs & cautions
Not a complete database: access control, metadata, persistence and updates need design.
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 / self-hosted
Lifecycle
Not independently audited
Comparison basis
Editorial problem-fit judgement; no head-to-head benchmark performed
Connected concepts
Alternatives to investigate
Qdrant · hnswlib · pgvector