• Flat (brute force)
• IVF — Inverted File Index
• HNSW — Hierarchical Navigable Small World
• PQ — Product Quantization (and the IVF-PQ combo)
• ScaNN — Google's, used in Vertex AI Vector Search
• DiskANN — when you need scale beyond RAM

The mental cheat sheet

• Small corpus (<100K): flat is fine, don't overthink it
• Medium (100K - 10M), latency-critical, RAM-heavy: HNSW
• Large (10M - 100M), memory-budget-conscious: IVF-PQ or ScaNN
• Massive (>100M, billions): ScaNN, DiskANN, or sharded HNSW
• Need fast updates on a small corpus: flat or HNSW with relaxed M
• Static corpus, build once query forever: anything's fine, pick by recall/latency