Vector Index Tuning

Benchmarks HNSW parameters (M, efConstruction, efSearch) against your actual data to find the sweet spot between recall and latency. Implements scalar quantization, product quantization, and binary compression with memory usage estimates. The parameter tuning function runs systematic tests across different configurations and calculates recall@k, which beats guessing at settings. Also includes quantization strategies that can cut memory usage by 75% with minimal recall loss. Most useful when you're hitting memory limits or search latency SLAs with large vector datasets over 100K embeddings.

skills.sh installs: 11K
GitHub stars: 42.4K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add wshobson/agents --skill vector-index-tuning --agent claude-code

Installs into .claude/skills of the current project.

Files

Vector Index Tuning

Guide to optimizing vector indexes for production performance.

When to Use This Skill

Tuning HNSW parameters
Implementing quantization
Optimizing memory usage
Reducing search latency
Balancing recall vs speed
Scaling to billions of vectors

Core Concepts

1. Index Type Selection

Data Size           Recommended Index
────────────────────────────────────────
< 10K vectors  →    Flat (exact search)
10K - 1M       →    HNSW
1M - 100M      →    HNSW + Quantization
> 100M         →    IVF + PQ or DiskANN

2. HNSW Parameters

Parameter	Default	Effect
M	16	Connections per node, ↑ = better recall, more memory
efConstruction	100	Build quality, ↑ = better index, slower build
efSearch	50	Search quality, ↑ = better recall, slower search